Visit https://ebookultra.

com to download the full version and

explore more ebooks

Computation and Modelling in Insurance and Finance

Boelviken E.

_____ Click the link below to download _____

https://ebookultra.com/download/computation-and-
modelling-in-insurance-and-finance-boelviken-e/

Explore and download more ebooks at ebookultra.com

Here are some recommended products that might interest you.
You can download now and explore!

Risk Analysis in Finance and Insurance Second Edition

Alexander Melnikov

https://ebookultra.com/download/risk-analysis-in-finance-and-
insurance-second-edition-alexander-melnikov/

ebookultra.com

Career Opportunities in Banking Finance And Insurance 2nd

ed Edition Thomas P. Fitch

https://ebookultra.com/download/career-opportunities-in-banking-
finance-and-insurance-2nd-ed-edition-thomas-p-fitch/

ebookultra.com

Risk Modelling in General Insurance From Principles to

Practice 1st Edition Roger J. Gray

https://ebookultra.com/download/risk-modelling-in-general-insurance-
from-principles-to-practice-1st-edition-roger-j-gray/

ebookultra.com

Principles of Corporate Finance Mcgraw Hill Irwin Series

in Finance Insurance Real Estate 7th Edition Richard
Brealey
https://ebookultra.com/download/principles-of-corporate-finance-
mcgraw-hill-irwin-series-in-finance-insurance-real-estate-7th-edition-
richard-brealey/
ebookultra.com
Synthetic CDOs Modelling Valuation and Risk Management
Mathematics Finance and Risk 1st Edition C. C. Mounfield

https://ebookultra.com/download/synthetic-cdos-modelling-valuation-
and-risk-management-mathematics-finance-and-risk-1st-edition-c-c-
mounfield/
ebookultra.com

Insurance Statistics Yearbook 2008 Edition 2008 FINANCE ET

INVESTISSEMENT ASSURANCE ET Oecd

https://ebookultra.com/download/insurance-statistics-
yearbook-2008-edition-2008-finance-et-investissement-assurance-et-
oecd/
ebookultra.com

Interest Rates and Coupon Bonds in Quantum Finance 1st

Edition Belal E. Baaquie

https://ebookultra.com/download/interest-rates-and-coupon-bonds-in-
quantum-finance-1st-edition-belal-e-baaquie/

ebookultra.com

Modelling Single name and Multi name Credit Derivatives

The Wiley Finance Series 1st Edition Dominic O'Kane

https://ebookultra.com/download/modelling-single-name-and-multi-name-
credit-derivatives-the-wiley-finance-series-1st-edition-dominic-okane/

ebookultra.com

Careers in Insurance and Risk Management 2009 Edition

Wetfeet

https://ebookultra.com/download/careers-in-insurance-and-risk-
management-2009-edition-wetfeet/

ebookultra.com
Computation and Modelling in Insurance and Finance
Boelviken E. Digital Instant Download
Author(s): Boelviken E.
ISBN(s): 9780521830485, 0521830486
Edition: draft
File Details: PDF, 3.28 MB
Year: 2014
Language: english
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page i — #1

i i

Computation and Modelling in Insurance and Finance

Focusing on what actuaries need in practice, this introductory account provides readers with
essential tools for handling complex problems and explains how simulation models can be created,
used and reused (with modifications) in related situations.
The book begins by outlining the basic tools of modelling and simulation, including a discussion
of the Monte Carlo method and its use. Part II deals with general insurance and Part III with life
insurance and financial risk. Algorithms that can be implemented on any programming platform are
spread throughout, and a program library written in R is included. Numerous figures and
experiments with R-code illustrate the text.
The authors non-technical approach is ideal for graduate students, the only prerequisites being
introductory courses in calculus and linear algebra, probability and statistics. The book will also be
of value to actuaries and other analysts in the industry looking to update their skills.

e r i k b ø lv i k e n holds the position of Chair of Actuarial Science at the University of Oslo and is
a partner in Gabler and Partners, Oslo.

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page ii — #2

i i

INTERNATIONAL SERIES ON ACTUARIAL SCIENCE

Editorial Board
Christopher Daykin (Independent Consultant and Actuary)
Angus Macdonald (Heriot-Watt University)

The International Series on Actuarial Science, published by Cambridge University Press

in con-junction with the Institute and Faculty of Actuaries, contains textbooks for students
taking courses in or related to actuarial science, as well as more advanced works designed
for continuing pro-fessional development or for describing and synthesizing research. The
series is a vehicle for publishing books that reflect changes and developments in the cur-
riculum, that encourage the introduction of courses on actuarial science in universities, and
that show how actuarial science can be used in all areas where there is long-term financial
risk.
A complete list of books in the series can be found at www.cambridge.org/statistics. Recent
titles include the following:
Solutions Manual for Actuarial Mathematics for Life Contingent Risks (2nd Edition)
David C.M. Dickson, Mary R. Hardy & Howard R. Waters
Actuarial Mathematics for Life Contingent Risks (2nd Edition)
David C.M. Dickson, Mary R. Hardy & Howard R. Waters
Risk Modelling in General Insurance
Roger J. Gray & Susan M. Pitts
Financial Enterprise Risk Management
Paul Sweeting
Regression Modeling with Actuarial and Financial Applications
Edward W. Frees
Nonlife Actuarial Models
Yiu-Kuen Tse
Generalized Linear Models for Insurance Data
Piet De Jong & Gillian Z. Heller
Market-Valuation Methods in Life and Pension Insurance
Thomas Møller & Mogens Steffensen

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page iii — #3

i i

COMPUTATION AND MODELLING IN

INSURANCE AND FINANCE

E R I K B Ø LV I K E N
University of Oslo

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page iv — #4

i i

University Printing House, Cambridge CB2 8BS, United Kingdom

Published in the United States of America by Cambridge University Press, New York

Cambridge University Press is part of the University of Cambridge.

It furthers the University’s mission by disseminating knowledge in the pursuit of
education, learning and research at the highest international levels of excellence.

www.cambridge.org
Information on this title: www.cambridge.org/9780521830485

c Cambridge University Press 2014
This publication is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published 2014
Printing in the United Kingdom by TJ International Ltd. Padstow Cornwall
A catalogue record for this publication is available from the British Library
Library of Congress Cataloguing in Publication data
ISBN 978-0-521-83048-5 Hardback
Cambridge University Press has no responsibility for the persistence or accuracy of
URLs for external or third-party internet websites referred to in this publication,
and does not guarantee that any content on such websites is, or will remain,
accurate or appropriate.

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page v — #5

i i

Contents

Preface page xxv

1 Introduction 1
1.1 A view on the evaluation of risk 1
1.1.1 The role of mathematics 1
1.1.2 Risk methodology 1
1.1.3 The computer model 3
1.2 Insurance risk: Basic concepts 4
1.2.1 Introduction 4
1.2.2 Pricing insurance risk 4
1.2.3 Portfolios and solvency 5
1.2.4 Risk ceding and reinsurance 6
1.3 Financial risk: Basic concepts 6
1.3.1 Introduction 6
1.3.2 Rates of interest 7
1.3.3 Financial returns 7
1.3.4 Log-returns 8
1.3.5 Financial portfolios 8
1.4 Risk over time 9
1.4.1 Introduction 9
1.4.2 Accumulation of values 10
1.4.3 Forward rates of interest 10
1.4.4 Present and fair values 11
1.4.5 Bonds and yields 12
1.4.6 Duration 12
1.4.7 Investment strategies 13
1.5 Method: A unified beginning 14
1.5.1 Introduction 14

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page vi — #6

i i

vi Contents

1.5.2 Monte Carlo algorithms and notation 14

1.6 How the book is planned 19
1.6.1 Mathematical level 19
1.6.2 Organization 20
1.6.3 Exercises and R 20
1.6.4 Notational rules and conventions 21
1.7 Bibliographical notes 21
1.7.1 General work 21
1.7.2 Monte Carlo 22

PART I TOOLS FOR RISK ANALYSIS 23

2 Getting started the Monte Carlo way 25
2.1 Introduction 25
2.2 How simulations are used 26
2.2.1 Introduction 26
2.2.2 Mean and standard deviation 26
2.2.3 Example: Financial returns 26
2.2.4 Percentiles 28
2.2.5 Density estimation 28
2.2.6 Monte Carlo error and selection of m 30
2.3 How random variables are sampled 31
2.3.1 Introduction 31
2.3.2 Inversion 31
2.3.3 Acceptance–rejection 32
2.3.4 Ratio of uniforms 33
2.4 Making the Gaussian work 34
2.4.1 Introduction 34
2.4.2 The normal family 34
2.4.3 Modelling on logarithmic scale 35
2.4.4 Stochastic volatility 36
2.4.5 The t-family 37
2.4.6 Dependent normal pairs 38
2.4.7 Dependence and heavy tails 39
2.4.8 Equicorrelation models 40
2.5 Positive random variables 40
2.5.1 Introduction 40
2.5.2 The Gamma distribution 41
2.5.3 The exponential distribution 42

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page vii — #7

i i

Contents vii

2.5.4 The Weibull distribution 42

2.5.5 The Pareto distribution 43
2.5.6 The Poisson distribution 44
2.6 Mathematical arguments 45
2.6.1 Monte Carlo error and tails of distributions 45
2.6.2 Algorithm 2.7 revisited 46
2.6.3 Algorithm 2.9 revisited 46
2.6.4 Algorithm 2.10 revisited 47
2.6.5 Algorithm 2.14 revisited 47
2.7 Bibliographical notes 48
2.7.1 Statistics and distributions 48
2.7.2 Sampling 48
2.7.3 Programming 48
2.8 Exercises 49
3 Evaluating risk: A primer 61
3.1 Introduction 61
3.2 General insurance: Opening look 62
3.2.1 Introduction 62
3.2.2 Enter contracts and their clauses 62
3.2.3 Stochastic modelling 63
3.2.4 Risk diversification 64
3.3 How Monte Carlo is put to work 64
3.3.1 Introduction 64
3.3.2 Skeleton algorithms 65
3.3.3 Checking the program 65
3.3.4 Computing the reserve 66
3.3.5 When responsibility is limited 67
3.3.6 Dealing with reinsurance 68
3.4 Life insurance: A different story 69
3.4.1 Introduction 69
3.4.2 Life insurance uncertainty 69
3.4.3 Life insurance mathematics 70
3.4.4 Simulating pension schemes 71
3.4.5 Numerical example 72
3.5 Financial risk: Derivatives as safety 72
3.5.1 Introduction 72
3.5.2 Equity puts and calls 73
3.5.3 How equity options are valued 74
3.5.4 The Black–Scholes formula 75

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page viii — #8

i i

viii Contents

3.5.5 Options on portfolios 75

3.5.6 Are equity options expensive? 76
3.6 Risk over long terms 77
3.6.1 Introduction 77
3.6.2 The ruin problem 78
3.6.3 Cash flow simulations 78
3.6.4 Solving the ruin equation 79
3.6.5 An underwriter example 80
3.6.6 Financial income added 80
3.7 Mathematical arguments 82
3.7.1 The Black–Scholes formula 82
3.7.2 The derivative with respect to σ 83
3.7.3 Solvency without financial earning 83
3.7.4 A representation of net assets 83
3.8 Bibliographical notes 84
3.8.1 General work 84
3.8.2 Monte Carlo and implementation 84
3.8.3 Other numerical methods 84
3.9 Exercises 84
4 Monte Carlo II: Improving technique 97
4.1 Introduction 97
4.2 Table look-up methods 97
4.2.1 Introduction 97
4.2.2 Uniform sampling 98
4.2.3 General discrete sampling 98
4.2.4 Example: Poisson sampling 100
4.2.5 Making the continuous discrete 100
4.2.6 Example: Put options 101
4.3 Correlated sampling 102
4.3.1 Introduction 102
4.3.2 Common random numbers 102
4.3.3 Example from finance 104
4.3.4 Example from insurance 104
4.3.5 Negative correlation: Antithetic designs 105
4.3.6 Examples of designs 106
4.3.7 Antithetic design in property insurance 107
4.4 Importance sampling and rare events 108
4.4.1 Introduction 108
4.4.2 The sampling method 108

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page ix — #9

i i

Contents ix

4.4.3 Choice of importance distribution 109

4.4.4 Importance sampling in property insurance 109
4.4.5 Application: Reserves and reinsurance premia 110
4.4.6 Example: A Pareto portfolio 111
4.5 Control variables 112
4.5.1 Introduction 112
4.5.2 The control method and reinsurance 113
4.5.3 Control scheme with equity options 113
4.5.4 Example: Put options 114
4.6 Random numbers: Pseudo- and quasi- 115
4.6.1 Introduction 115
4.6.2 Pseudo-random numbers 115
4.6.3 Quasi-random numbers: Preliminaries 116
4.6.4 Sobol sequences: Construction 118
4.6.5 Higher dimension and random shifts 119
4.6.6 Quasi-random numbers: Accuracy 120
4.7 Mathematical arguments 121
4.7.1 Efficiency of antithetic designs 121
4.7.2 Antithetic variables and property insurance 121
4.7.3 Importance sampling 123
4.7.4 Control scheme for equity options 123
4.8 Bibliographical notes 124
4.8.1 General work 124
4.8.2 Special techniques 124
4.8.3 Markov chain Monte Carlo 125
4.8.4 High-dimensional systems 125
4.9 Exercises 126
5 Modelling I: Linear dependence 138
5.1 Introduction 138
5.2 Descriptions of first and second order 138
5.2.1 Introduction 138
5.2.2 What a correlation tells us 139
5.2.3 Many correlated variables 140
5.2.4 Estimation using historical data 141
5.3 Financial portfolios and Markowitz theory 142
5.3.1 Introduction 142
5.3.2 The Markowitz problem 142
5.3.3 Solutions 143
5.3.4 Numerical illustration 144

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page x — #10

i i

x Contents

5.3.5 Two risky assets 144

5.3.6 Example: The crash of a hedge fund 145
5.3.7 Diversification of financial risk I 146
5.3.8 Diversification under CAPM 146
5.4 Dependent Gaussian models once more 147
5.4.1 Introduction 147
5.4.2 Uniqueness 148
5.4.3 Properties 148
5.4.4 Simulation 149
5.4.5 Scale for modelling 149
5.4.6 Numerical example: Returns or log-returns? 150
5.4.7 Making tails heavier 150
5.5 The random walk 151
5.5.1 Introduction 151
5.5.2 Random walk and equity 151
5.5.3 Elementary properties 152
5.5.4 Several processes jointly 153
5.5.5 Simulating the random walk 153
5.5.6 Numerical illustration 154
5.6 Introducing stationary models 155
5.6.1 Introduction 155
5.6.2 Autocovariances and autocorrelations 155
5.6.3 Estimation from historical data 156
5.6.4 Autoregression of first order 157
5.6.5 The behaviour of first-order autoregressions 158
5.6.6 Non-linear change of scale 159
5.6.7 Monte Carlo implementation 160
5.6.8 Numerical illustration 160
5.7 Changing the time scale 161
5.7.1 Introduction 161
5.7.2 Historical data on short time scales 162
5.7.3 The random walk revisited 162
5.7.4 Continuous time: The Wiener process 163
5.7.5 First-order autoregression revisited 164
5.7.6 Continuous-time autoregression 165
5.8 Mathematical arguments 166
5.8.1 Markowitz optimality 166
5.8.2 Risk bound under CAPM 166
5.8.3 Covariances of first-order autoregressions 167
5.8.4 Volatility estimation and time scale 167

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page xi — #11

i i

Contents xi

5.8.5 The accuracy of covariance estimates 167

5.9 Bibliographical notes 168
5.9.1 General work 168
5.9.2 Continuous-time processes 169
5.9.3 Historical data and the time scale 169
5.10 Exercises 169
6 Modelling II: Conditional and non-linear 182
6.1 Introduction 182
6.2 Conditional modelling 183
6.2.1 Introduction 183
6.2.2 The conditional Gaussian 183
6.2.3 Survival modelling 184
6.2.4 Over-threshold modelling 184
6.2.5 Stochastic parameters 185
6.2.6 Common factors 185
6.2.7 Monte Carlo distributions 186
6.3 Uncertainty on different levels 187
6.3.1 Introduction 187
6.3.2 The double rules 187
6.3.3 Financial risk under CAPM 188
6.3.4 Insurance risk 188
6.3.5 Impact of subordinate risk 189
6.3.6 Random claim intensity in general insurance 190
6.4 The role of the conditional mean 191
6.4.1 Introduction 191
6.4.2 Optimal prediction and interest rates 191
6.4.3 The conditional mean as a price 192
6.4.4 Modelling bond prices 192
6.4.5 Bond price schemes 193
6.4.6 Interest rate curves 194
6.5 Stochastic dependence: General 195
6.5.1 Introduction 195
6.5.2 Factorization of density functions 195
6.5.3 Types of dependence 196
6.5.4 Linear and normal processes 197
6.5.5 The multinomial situation 197
6.6 Markov chains and life insurance 198
6.6.1 Introduction 198
6.6.2 Markov modelling 199

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page xii — #12

i i

xii Contents

6.6.3 A disability scheme 200

6.6.4 Numerical example 200
6.7 Introducing copulas 202
6.7.1 Introduction 202
6.7.2 Copula modelling 202
6.7.3 The Clayton copula 203
6.7.4 Conditional distributions under copulas 204
6.7.5 Many variables and the Archimedean class 205
6.7.6 The Marshall–Olkin representation 206
6.7.7 Copula sampling 207
6.7.8 Example: An equity portfolio 208
6.7.9 Example: Copula log-normals against pure log-normals 209
6.8 Mathematical arguments 210
6.8.1 Portfolio risk when claim intensities are random 210
6.8.2 Optimal prediction 211
6.8.3 Vasic̆ek bond prices 211
6.8.4 The Marshall–Olkin representation 212
6.8.5 A general scheme for copula sampling 213
6.8.6 Justification of Algorithm 6.4 213
6.9 Bibliographical notes 214
6.9.1 General work 214
6.9.2 Applications 214
6.9.3 Copulas 214
6.10 Exercises 215
7 Historical estimation and error 229
7.1 Introduction 229
7.2 Error of different origin 230
7.2.1 Introduction 230
7.2.2 Quantifying error 231
7.2.3 Numerical illustration 232
7.2.4 Errors and the mean 232
7.2.5 Example: Option pricing 233
7.2.6 Example: Reserve in property insurance 234
7.2.7 Bias and model error 235
7.3 How parameters are estimated 236
7.3.1 Introduction 236
7.3.2 The quick way: Moment matching 237
7.3.3 Moment matching and time series 237
7.3.4 The usual way: Maximum likelihood 238

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page xiii — #13

i i

Contents xiii

7.3.5 Example: Norwegian natural disasters 239

7.4 Evaluating error I 240
7.4.1 Introduction 240
7.4.2 Introducing the bootstrap 241
7.4.3 Introductory example: The Poisson bootstrap 242
7.4.4 Second example: The Pareto bootstrap 243
7.4.5 The pure premium bootstrap 244
7.4.6 Simplification: The Gaussian bootstrap 245
7.4.7 The old way: Delta approximations 246
7.5 Evaluating error II: Nested schemes 247
7.5.1 Introduction 247
7.5.2 The nested algorithm 247
7.5.3 Example: The reserve bootstrap 248
7.5.4 Numerical illustration 248
7.5.5 A second example: Interest-rate return 249
7.5.6 Numerical illustration 251
7.6 The Bayesian approach 252
7.6.1 Introduction 252
7.6.2 The posterior view 252
7.6.3 Example: Claim intensities 253
7.6.4 Example: Expected return on equity 253
7.6.5 Choosing the prior 255
7.6.6 Bayesian simulation 255
7.6.7 Example: Mean payment 256
7.6.8 Example: Pure premium 258
7.6.9 Summing up: Bayes or not? 258
7.7 Mathematical arguments 258
7.7.1 Bayesian means under Gaussian models 258
7.8 Bibliographical notes 259
7.8.1 Analysis of error 259
7.8.2 The bootstrap 259
7.8.3 Bayesian techniques 260
7.9 Exercises 260

PART II GENERAL INSURANCE 277

8 Modelling claim frequency 279
8.1 Introduction 279
8.2 The world of Poisson 279

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page xiv — #14

i i

xiv Contents

8.2.1 Introduction 279

8.2.2 An elementary look 280
8.2.3 Extending the argument 280
8.2.4 When the intensity varies over time 282
8.2.5 The Poisson distribution 283
8.2.6 Using historical data 283
8.2.7 Example: A Norwegian automobile portfolio 284
8.3 Random intensities 285
8.3.1 Introduction 285
8.3.2 A first look 286
8.3.3 Estimating the mean and variance of μ 286
8.3.4 The negative binomial model 287
8.3.5 Fitting the negative binomial 288
8.3.6 Automobile example continued 288
8.4 Intensities with explanatory variables 289
8.4.1 Introduction 289
8.4.2 The model 289
8.4.3 Data and likelihood function 289
8.4.4 A first interpretation 290
8.4.5 How variables are entered 291
8.4.6 Interaction and cross-classification 292
8.5 Modelling delay 293
8.5.1 Introduction 293
8.5.2 Multinomial delay 294
8.5.3 IBNR claim numbers 294
8.5.4 Fitting delay models 295
8.5.5 Syntetic example: Car crash injury 295
8.6 Mathematical arguments 297
8.6.1 The general Poisson argument 297
8.6.2 Estimating the mean and standard deviation of μ 298
8.6.3 Large-sample properties 299
8.6.4 The negative binomial density function 300
8.6.5 Skewness of the negative binomial 301
8.6.6 The convolution property of the negative binomial 301
8.6.7 IBNR: The delay model 302
8.7 Bibliographical notes 302
8.7.1 Poisson modelling 302
8.7.2 Generalized linear models 302
8.7.3 Reserving over long 303
8.8 Exercises 303

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page xv — #15

i i

Contents xv

9 Modelling claim size 314

9.1 Introduction 314
9.2 Parametric and non-parametric modelling 314
9.2.1 Introduction 314
9.2.2 Scale families of distributions 315
9.2.3 Fitting a scale family 316
9.2.4 Shifted distributions 316
9.2.5 Skewness as a simple description of shape 317
9.2.6 Non-parametric estimation 318
9.3 The log-normal and Gamma families 319
9.3.1 Introduction 319
9.3.2 The log-normal: A quick summary 319
9.3.3 The Gamma model 319
9.3.4 Fitting the Gamma familiy 320
9.3.5 Regression for claims size 321
9.4 The Pareto families 322
9.4.1 Introduction 322
9.4.2 Elementary properties 322
9.4.3 Likelihood estimation 323
9.4.4 Over-threshold under Pareto 323
9.4.5 The extended Pareto family 324
9.5 Extreme value methods 325
9.5.1 Introduction 325
9.5.2 Over-threshold distributions in general 325
9.5.3 The Hill estimate 327
9.5.4 The entire distribution through mixtures 327
9.5.5 The empirical distribution mixed with Pareto 328
9.6 Searching for the model 329
9.6.1 Introduction 329
9.6.2 Using transformations 329
9.6.3 Example: The Danish fire claims 330
9.6.4 Pareto mixing 331
9.6.5 When data are scarce 332
9.6.6 When data are scarce II 333
9.7 Mathematical arguments 334
9.7.1 Extended Pareto: Moments 334
9.7.2 Extended Pareto: A limit 335
9.7.3 Extended Pareto: A representation 335
9.7.4 Extended Pareto: Additional sampler 336
9.7.5 Justification of the Hill estimate 336

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page xvi — #16

i i

xvi Contents

9.8 Bibliographical notes 337

9.8.1 Families of distributions 337
9.8.2 Extreme value theory 337
9.8.3 Over thresholds 337
9.9 Exercises 337
10 Solvency and pricing 351
10.1 Introduction 351
10.2 Portfolio liabilities by simple approximation 352
10.2.1 Introduction 352
10.2.2 Normal approximations 352
10.2.3 Example: Motor insurance 353
10.2.4 The normal power approximation 353
10.2.5 Example: Danish fire claims 354
10.3 Portfolio liabilities by simulation 355
10.3.1 Introduction 355
10.3.2 A skeleton algorithm 355
10.3.3 Danish fire data: The impact of the claim size model 356
10.4 Differentiated pricing through regression 357
10.4.1 Introduction 357
10.4.2 Estimates of pure premia 358
10.4.3 Pure premia regression in practice 359
10.4.4 Example: The Norwegian automobile portfolio 359
10.5 Differentiated pricing through credibility 361
10.5.1 Introduction 361
10.5.2 Credibility: Approach 361
10.5.3 Linear credibility 362
10.5.4 How accurate is linear credibility? 363
10.5.5 Credibility at group level 364
10.5.6 Optimal credibility 364
10.5.7 Estimating the structural parameters 365
10.6 Reinsurance 366
10.6.1 Introduction 366
10.6.2 Traditional contracts 366
10.6.3 Pricing reinsurance 367
10.6.4 The effect of inflation 368
10.6.5 The effect of reinsurance on the reserve 369
10.7 Mathematical arguments 370
10.7.1 The normal power approximation 370
10.7.2 The third-order moment of X 370

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page xvii — #17

i i

Contents xvii

10.7.3 Normal power under heterogeneity 371

10.7.4 Auxiliary for linear credibility 372
10.7.5 Linear credibility 372
10.7.6 Optimal credibility 373
10.7.7 Parameter estimation in linear credibility 374
10.8 Bibliographical notes 375
10.8.1 Computational methods 375
10.8.2 Credibility 375
10.8.3 Reinsurance 376
10.9 Exercises 376
11 Liabilities over long terms 386
11.1 Introduction 386
11.2 Simple situations 387
11.2.1 Introduction 387
11.2.2 Lower-order moments 387
11.2.3 When risk is constant 388
11.2.4 Underwriter results in the long run 388
11.2.5 Underwriter ruin by closed mathematics 389
11.2.6 Underwriter ruin under heavy-tailed losses 391
11.3 Time variation through regression 391
11.3.1 Introduction 391
11.3.2 Poisson regression with time effects 391
11.3.3 Example: An automobile portfolio 392
11.3.4 Regression with random background 393
11.3.5 The automobile portfolio: A second round 394
11.4 Claims as a stochastic process 396
11.4.1 Introduction 396
11.4.2 Claim intensity as a stationary process 396
11.4.3 A more general viewpoint 397
11.4.4 Model for the claim numbers 397
11.4.5 Example: The effect on underwriter risk 399
11.4.6 Utilizing historical data 399
11.4.7 Numerical experiment 400
11.5 Building simulation models 401
11.5.1 Introduction 401
11.5.2 Under the top level 401
11.5.3 Hidden, seasonal risk 402
11.5.4 Hidden risk with inflation 403
11.5.5 Example: Is inflation important? 404

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page xviii — #18

i i

xviii Contents

11.5.6 Market fluctuations 405

11.5.7 Market fluctuations: Example 406
11.5.8 Taxes and dividend 407
11.6 Cash flow or book value? 408
11.6.1 Introduction 408
11.6.2 Mathematical formulation 408
11.6.3 Adding IBNR claims 409
11.6.4 Example: Runoff portfolios 410
11.7 Mathematical arguments 411
11.7.1 Lower-order moments of Yk under constant risk 411
11.7.2 Lundberg’s inequality 412
11.7.3 Moment-generating functions for underwriting 412
11.7.4 Negative binomial regression 413
11.8 Bibliographical notes 414
11.8.1 Negative binomial regression 414
11.8.2 Claims as stochastic processes 414
11.8.3 Ruin 415
11.9 Exercises 415

PART III LIFE INSURANCE AND FINANCIAL RISK 431

12 Life and state-dependent insurance 433
12.1 Introduction 433
12.2 The anatomy of state-dependent insurance 434
12.2.1 Introduction 434
12.2.2 Cash flows determined by states 434
12.2.3 Equivalence pricing 435
12.2.4 The reserve 436
12.2.5 The portfolio viewpoint 437
12.3 Survival modelling 438
12.3.1 Introduction 438
12.3.2 Deductions from one-step transitions 438
12.3.3 Modelling through intensities 439
12.3.4 A standard model: Gomperz–Makeham 440
12.3.5 Expected survival 441
12.3.6 Using historical data 442
12.4 Single-life arrangements 443
12.4.1 Introduction 443
12.4.2 How mortality risk affects value 443

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page xix — #19

i i

Contents xix

12.4.3 Life insurance notation 444

12.4.4 Computing mortality-adjusted annuities 445
12.4.5 Common insurance arrangements 445
12.4.6 A numerical example 448
12.5 Multi-state insurance I: Modelling 448
12.5.1 Introduction 448
12.5.2 From one-step to k-step transitions 449
12.5.3 Intensity modelling 450
12.5.4 Example: A Danish disability model 451
12.5.5 Numerical examples 452
12.5.6 From intensities to transition probabilities 453
12.5.7 Using historical data 454
12.6 Multi-state insurance II: Premia and liabilities 455
12.6.1 Introduction 455
12.6.2 Single policies 455
12.6.3 Example 1: A widow scheme 455
12.6.4 Example 2: Disability and retirement in combination 456
12.6.5 Portfolio liabilities 457
12.6.6 Example: A disability scheme 458
12.7 Mathematical arguments 459
12.7.1 Savings and mortality-adjusted value 459
12.7.2 The reserve formula (12.19) 460
12.7.3 The k-step transition probabilities 461
12.8 Bibliographical notes 461
12.8.1 Simple contracts and modelling 461
12.8.2 General work 462
12.9 Exercises 462
13 Stochastic asset models 478
13.1 Introduction 478
13.2 Volatility modelling I 479
13.2.1 Introduction 479
13.2.2 Multivariate stochastic volatility 479
13.2.3 The multivariate t-distribution 480
13.2.4 Dynamic volatility 482
13.2.5 Volatility as driver 483
13.2.6 Log-normal volatility 483
13.2.7 Numerical example 484
13.2.8 Several series in parallel 485
13.3 Volatility modelling II: The GARCH type 485

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page xx — #20

i i

xx Contents

13.3.1 Introduction 485

13.3.2 How GARCH models are constructed 486
13.3.3 Volatilities under first-order GARCH 486
13.3.4 Properties of the original process 487
13.3.5 Fitting GARCH models 488
13.3.6 Simulating GARCH 488
13.3.7 Example: GARCH and the SP 500 index 489
13.4 Linear dynamic modelling 489
13.4.1 Introduction 489
13.4.2 ARMA models 491
13.4.3 Linear feedback 492
13.4.4 Enter transformations 493
13.5 The Wilkie model I: Twentieth-century financial risk 494
13.5.1 Introduction 494
13.5.2 Output variables and their building blocks 494
13.5.3 Non-linear transformations 496
13.5.4 The linear and stationary part 497
13.5.5 Parameter estimates 498
13.5.6 Annual inflation and returns 499
13.6 The Wilkie model II: Implementation issues 500
13.6.1 Introduction 500
13.6.2 How simulations are initialized 500
13.6.3 Simulation algorithms 502
13.6.4 Interest-rate interpolation 503
13.7 Mathematical arguments 503
13.7.1 Absolute deviations from the mean 503
13.7.2 Autocorrelations under log-normal volatilities 504
13.7.3 The error series for GARCH variances 505
13.7.4 Properties of GARCH variances 505
13.7.5 The original process squared 506
13.7.6 Verification of Table 13.5 507
13.8 Bibliographical notes 508
13.8.1 Linear processes 508
13.8.2 Heavy tails 508
13.8.3 Dynamic volatility 509
13.9 Exercises 509
14 Financial derivatives 522
14.1 Introduction 522
14.2 Arbitrage and risk neutrality 523

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page xxi — #21

i i

Contents xxi

14.2.1 Introduction 523

14.2.2 Forward contracts 524
14.2.3 Binomial movements 524
14.2.4 Risk neutrality 525
14.3 Equity options I 526
14.3.1 Introduction 526
14.3.2 Types of contract 527
14.3.3 Valuation: A first look 528
14.3.4 The put–call parity 528
14.3.5 A first look at calls and puts 529
14.4 Equity options II: Hedging and valuation 530
14.4.1 Introduction 530
14.4.2 Actuarial and risk-neutral pricing 530
14.4.3 The hedge portfolio and its properties 531
14.4.4 The financial state over time 533
14.4.5 Numerical experiment 533
14.4.6 The situation at expiry revisited 535
14.4.7 Valuation 536
14.5 Interest-rate derivatives 537
14.5.1 Introduction 537
14.5.2 Risk-neutral pricing 537
14.5.3 Implied mean and forward prices 538
14.5.4 Interest-rate swaps 539
14.5.5 Floors and caps 540
14.5.6 Options on bonds 541
14.5.7 Options on interest-rate swaps 542
14.5.8 Numerical experimenting 543
14.6 Mathematical summing up 543
14.6.1 Introduction 543
14.6.2 How values of derivatives evolve 545
14.6.3 Hedging 545
14.6.4 The market price of risk 546
14.6.5 Martingale pricing 547
14.6.6 Closing mathematics 548
14.7 Bibliographical notes 550
14.7.1 Introductory work 550
14.7.2 Work with heavier mathematics 550
14.7.3 The numerical side 550
14.8 Exercises 551

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page xxii — #22

i i

xxii Contents

15 Integrating risk of different origin 562

15.1 Introduction 562
15.2 Life-table risk 563
15.2.1 Introduction 563
15.2.2 Numerical example 564
15.2.3 The life-table bootstrap 565
15.2.4 The bootstrap in life insurance 566
15.2.5 Random error and pension evaluations 566
15.2.6 Bias against random error 568
15.2.7 Dealing with longer lives 569
15.2.8 Longevity bias: Numerical examples 570
15.3 Risk due to discounting and inflation 571
15.3.1 Introduction 571
15.3.2 Market-based valuation 572
15.3.3 Numerical example 572
15.3.4 Inflation: A first look 573
15.3.5 Simulating present values under stochastic discounts 574
15.3.6 Numerical examples 575
15.4 Simulating assets protected by derivatives 576
15.4.1 Introduction 576
15.4.2 Equity returns with options 577
15.4.3 Equity options over longer time horizons 577
15.4.4 Money-market investments with floors and caps 579
15.4.5 Money-market investments with swaps and swaptions 579
15.5 Simulating asset portfolios 581
15.5.1 Introduction 581
15.5.2 Defining the strategy 582
15.5.3 Expenses due to rebalancing 583
15.5.4 A skeleton algorithm 584
15.5.5 Example 1: Equity and cash 585
15.5.6 Example 2: Options added 586
15.5.7 Example 3: Bond portfolio and inflation 587
15.5.8 Example 4: Equity, cash and bonds 588
15.6 Assets and liabilities 589
15.6.1 Introduction 589
15.6.2 Auxiliary: Duration and spread of liabilities 590
15.6.3 Classical immunization 591
15.6.4 Net asset values 592
15.6.5 Immunization through bonds 593
15.6.6 Enter inflation 594

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page xxiii — #23

i i

Contents xxiii

15.7 Mathematical arguments 597

15.7.1 Present values and duration 597
15.7.2 Reddington immunization 597
15.8 Bibliographical notes 599
15.8.1 Survival modelling 599
15.8.2 Fair values 599
15.8.3 Financial risk and ALM 599
15.8.4 Stochastic dynamic optimization 600
15.9 Exercises 600
Appendix A Random variables: Principal tools 618
A.1 Introduction 618
A.2 Single random variables 618
A.2.1 Introduction 618
A.2.2 Probability distributions 618
A.2.3 Simplified description of distributions 619
A.2.4 Operating rules 620
A.2.5 Transforms and cumulants 620
A.2.6 Example: The mean 622
A.3 Several random variables jointly 622
A.3.1 Introduction 622
A.3.2 Covariance and correlation 623
A.3.3 Operating rules 624
A.3.4 The conditional viewpoint 624
A.4 Laws of large numbers 626
A.4.1 Introduction 626
A.4.2 The weak law of large numbers 626
A.4.3 Central limit theorem 627
A.4.4 Functions of averages 628
A.4.5 Bias and standard deviation of estimates 629
A.4.6 Likelihood estimates 629
Appendix B Linear algebra and stochastic vectors 631
B.1 Introduction 631
B.2 Operations on matrices and vectors 632
B.2.1 Introduction 632
B.2.2 Addition and multiplication 632
B.2.3 Quadratic matrices 632
B.2.4 The geometric view 633
B.2.5 Algebraic rules 634
B.2.6 Stochastic vectors 634

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page xxiv — #24

i i

xxiv Contents

B.2.7 Linear operations on stochastic vectors 635

B.2.8 Covariance matrices and Cholesky factors 636
B.3 The Gaussian model: Simple theory 636
B.3.1 Introduction 636
B.3.2 Orthonormal operations 637
B.3.3 Uniqueness 638
B.3.4 Linear transformations 638
B.3.5 Block representation 638
B.3.6 Conditional distributions 639
B.3.7 Verfication 639
Appendix C Numerical algorithms: A third tool 641
C.1 Introduction 641
C.2 Cholesky computing 641
C.2.1 Introduction 641
C.2.2 The Cholesky decomposition 642
C.2.3 Linear equations 642
C.2.4 Matrix inversion 643
C.3 Interpolation, integration, differentiation 644
C.3.1 Introduction 644
C.3.2 Numerical interpolation 644
C.3.3 Numerical integration 645
C.3.4 Numerical integration II: Gaussian quadrature 646
C.3.5 Numerical differentiation 647
C.4 Bracketing and bisection: Easy and safe 649
C.4.1 Introduction 649
C.4.2 Bisection: Bracketing as iteration 649
C.4.3 Golden section: Bracketing for extrema 650
C.4.4 Golden section: Justification 651
C.5 Optimization: Advanced and useful 652
C.5.1 Introduction 652
C.5.2 The Newton–Raphson method 652
C.5.3 Variable metric methods 653
C.6 Bibliographical notes 655
C.6.1 General numerical methods 655
C.6.2 Optimization 655
References 656
Index 680

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page xxv — #25

i i

Preface

The book is organized as a broad introduction to concepts, models and computa-

tional techniques in Part I and with general insurance and life insurance/financial
risk in Parts II and III. The latter are largely self-contained and can probably be read
on their own. Each part may be used as a basis for a university course; we do that
in Oslo. Computation is more strongly emphasized than in traditional textbooks.
Stochastic models are defined in the way they are simulated in the computer and
examined through numerical experiments. This cuts down on the mathematics and
enables students to reach ‘advanced’ models quickly. Numerical experimentation
is also a way to illustrate risk concepts and to indicate the impact of assumptions
that are often somewhat arbitrary. One of the aims of this book is to teach how the
computer is put to work effectively.
Other issues are error in risk assessments and the use of historical data, each of
which are tasks for statistics. Many of the models and distributions are presented
with simple fitting procedures, and there is an entire chapter on error analysis and
on the difference between risk under the underlying, real model and the one we
actually use. Such error is in my opinion often treated too lightly: we should be very
much aware of the distinction between the complex, random mechanisms in real
life and our simplified model versions with deviating parameters. In a nebulous and
ever-changing world modelling should be kept simple and limited to the essential.
The reader must be familiar with elementary calculus, probability and matrix
algebra (the last two being reviewed in appendices) and should preferably have
some programming experience. These apart, the book is self-contained with con-
cepts and methods developed from scratch. The text is equipped with algorithms
written in pseudo-code that can be programmed on any platform whereas the exer-
cises make use of the open-source R software which permits Monte Carlo simula-
tion of quite complex problems through a handful of commands. People can teach
themselves the tricks of R programming by following the instructions in the exer-
cises (my recommendation), but it is also possible to use the associated R library

xxv

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page xxvi — #26

i i

xxvi Preface

passively. The exercises vary from the theoretical to the numerical. There is a good
deal of experimentation and model comparison.
I am grateful to my editor at Cambridge University Press, David Tranah, for
proposing the project and for following the work with infinite patience over a much
longer period than originally envisaged. Montserrat Guillen provided a stimulating
environment at The University of Barcelona where some parts were written. Ragnar
Norberg at London School of Economics was kind enough to go through an earlier
version of the entire manuscript. My friends at GablerWassum in Oslo, above all
Christian Fotland and Arve Moe have been a source of inspiration. Jostein Sørvoll,
then head of the same company, taught me some of the practical sides of insurance
as did Torbjørn Jakobsen for finance. I also thank Morten Folkeson and Steinar
Holm for providing some of the historical data. Many of the exercises were checked
by my students Eirik Sagstuen and Rebecca Wiborg.

Erik Bølviken
Oslo

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page 1 — #27

i i

1
Introduction

1.1 A view on the evaluation of risk

1.1.1 The role of mathematics
How is evaluation of risk influenced by modern computing? Consider the way we
use mathematics, first as a vendor of models of complicated risk processes. These
models are usually stochastic. They are in general, insurance probability distribu-
tions of claim numbers and losses and in life insurance and finance, stochastic pro-
cesses describing lifecycles and investment returns. Mathematics is from this point
of view a language, a way risk is expressed, and it is a language we must master.
Otherwise statements of risk cannot be related to reality, it would be impossible to
say what conclusions mean in any precise manner and nor could analyses be pre-
sented effectively to clients. Actuarial science is in this sense almost untouched by
modern computational facilities. The basic concepts and models remain what they
were, notwithstanding, of course, the strong growth of risk products throughout the
last decades. This development may have had something to do with computers, but
not much with computing per se.
However, mathematics is also deductions with precise conclusions derived from
precise assumptions through the rules of logic. That is the way mathematics is
taught at school and university. It is here that computing enters applied mathemati-
cal disciplines like actuarial science. More and more of these deductions are imple-
mented in computers and carried out there. This has been going on for decades. It
leans on an endless growth in computing power, a true technological revolution
opening up simpler and more general computational methods which require less of
users.

1.1.2 Risk methodology

An example of such an all-purpose computational technique is stochastic simula-
tion. Simplified versions of processes taking place in real life are then reproduced

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page 2 — #28

i i

2 Introduction

The world In the computer 

Sources for M:
of risk Historical experience
M→X  → X∗
M The implied market view
↑ Deductions from no-arbitrage
Assumed mechanism Judgement, physical modelling

Figure 1.1 The working process: Main steps in risk evaluation.

in the computer. Risk in finance and insurance is future uncertain gains and losses,
designated in this book by letters such as X and Y. Typical examples are compen-
sations for claims in general insurance, pension schemes interrupted upon death
in life insurance and future values of shares and bonds in finance. There are also
secondary (or derived) products where values and payoffs are channelled through
contract clauses set up in advance. Such agreements are known as derivatives in
finance and reinsurance in insurance.
The mathematical approach, unanimously accepted today, is through probabilities
with risks X and Y seen as random variables. We shall know their values eventually
(after the event), but for planning and control and to price risk-taking activities
we need them in advance and must fall back on their probabilities. This leads to a
working process such as the one depicted in Figure 1.1. The real world on the left is
an enormously complicated mechanism (denoted M) that yields a future X.
We shall never know M, though our paradigm is that it does exist as a well-
defined stochastic mechanism. Since it is beyond reach, a simplified version M  is
constructed in its place and used to study X. Its expected value is used for valuation
and the percentiles for control, and unlike in engineering we are rarely concerned
with predicting a specific value. Note that everything falls apart if M  deviates too
strongly from the true mechanism M. This issue of error is a serious one indeed.
Chapter 7 is an introduction.
What there is to go on when M  is constructed is listed on the right in Figure 1.1.
Learning from the past is an obvious source (but not all of it is relevant). In finance,
current asset prices bring market opinion about the future. This so-called implied
view is introduced briefly in Section 1.4, and there will be more in Part III. Then
there is the theory of arbitrage, where riskless financial income is assumed impos-
sible. The innocent-looking no-arbitrage condition has wide implications, which
are discussed in Chapter 14. In practice, some personal judgement behind M  is
often present, but this is not for general argument, and nor shall we go into the
physical modelling used in large-claims insurance where hurricanes, earthquakes
or floods are imitated in the computer. This book is about how M  is constructed
from the first three sources (historical data above all), how it is implemented in
the computer and how the computer model is used to determine the probability

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page 3 — #29

i i

1.1 A view on the evaluation of risk 3

 is inevitably linked to the past even though perceived

distribution of X. Note that M
trends and changes may have been built into it, and there is no way of knowing how
well it captures a future which sometimes extends over decades. While this doesn’t
make the mathematical approach powerless, it does suggest simple and transparent
models and a humble attitude towards it all.

1.1.3 The computer model

The real risk variable X will materialize only once. The economic result of a finan-
cial investment in a particular year is a unique event, as is the aggregated claim
against an insurance portfolio during a certain period of time. With the computer
model, that is different. Once it has been set up it can be played as many times as
we please. Let X1∗ , . . . , Xm∗ be realizations of X revealing which values are likely
and which are not, and how bad things might be if we are unlucky. The ∗ will be
used to distinguish computer simulations from real variables and m will always
denote the number of simulations.
The method portrayed on the left of Figure 1.1 is known as the Monte Carlo
method or stochastic simulation. It belongs to the realm of numerical integration;
see Evans and Schwarz (2000) for a summary of this important branch of numerical
mathematics. Monte Carlo integration dates back a long way. It is computationally
slow, but other numerical methods (that might do the job faster) often require more
expertise and get bogged down for high-dimensional integrals, which are precisely
what we often need in practice. The Monte Carlo method is unique in handling
many variables well.
What is the significance of numerical speed anyway? Does it really matter that
some specialized technique (demanding more time and know-how to implement)
is (say) one hundred times faster when the one we use only takes a second? If
the procedure for some reason is to be repeated in a loop thousands of times, it
would matter. Often, however, slow Monte Carlo is quite enough, and, indeed, the
practical limit to its use is moving steadily as computers become more and more
powerful. How far have we got? The author’s portable computer from 2006 (with
T60p processor) took three seconds to produce ten million Pareto draws through
Algorithm 2.13 implemented in Fortran. This is an insurance portfolio of 1000
claims simulated 10 000 times! Generating the normal is even faster, and if speed
is a priority, try the table methods in Section 4.2.
One of the aims of this book is to demonstrate how these opportunities are uti-
lized. Principal issues are how simulations programs are designed, how they are
modified to deal with related (but different) problems and how different programs
are merged to handle situations of increasing complexity with several risk factors
contributing jointly. The versatility and usefulness of Monte Carlo is indicated in

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page 4 — #30

i i

4 Introduction

Section 1.5 (and in Chapter 3 too). By mastering it you are well equipped to deal
with much that comes your way and avoid getting stuck when pre-programmed
software doesn’t have what you need. What platform should you go for? Algo-
rithms in this book are written in a pseudo-code that goes with everything. Excel
and Visual Basic are standard in the industry and may be used even for simulation.
Much higher speed is obtained with C, Pascal or Fortran, and in the opinion of this
author people are well advised to learn software like those. There are other possi-
bilities as well, and the open-source R-package is used with the exercises. Much
can be achieved with a platform you know!

1.2 Insurance risk: Basic concepts

1.2.1 Introduction
Property or general insurance is economic responsibility for incidents such as
fires or accidents passed on (entirely or in part) to an insurer against a fee. The
contract, known as a policy, releases indemnities (claims) when such events occur.
A central quantity is the total claim X amassed during a certain period of time
(typically a year). Often X = 0 (no events), but on rare occasions X is huge. An
insurance company copes whatever happens, if properly run. It has a portfolio of
many such risks and only a few of them materialize. But this raises the issue of
controlling the total uncertainty, which is a major theme in general insurance.
Life insurance is also built up from random payments X. Term insurance,
where beneficiaries receive compensation upon the death of the policy holder, is
similar to property insurance in that unexpected events lead to payoffs. Pension
schemes are the opposite. Now the payments go on as long as the insured is alive,
and they are likely, not rare. Yet the basic approach remains the same, with random
variables X expressing the uncertainty involved.

1.2.2 Pricing insurance risk

Transfers of risk through X do not take place for free. The fee (or premium),
charged in advance, depends on the market conditions, but the expectation is a
guideline. Introduce

πpu = E(X), (1.1)

which is known as the pure premium and defines a break-even situation. A com-
pany receiving πpu for its services will, in the absence of all overhead cost and all
financial income, neither earn nor lose in the long run. This is a consequence of the
law of large numbers in probability theory; see Appendix A.2.

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page 5 — #31

i i

1.2 Insurance risk: Basic concepts 5

Such a pricing strategy is (of course) out of the question, and companies add
loadings γ on top of πpu . The premium charged is then
π = (1 + γ)πpu , (1.2)
and we may regard γπpu as the cost of risk. It is influenced thoroughly by the
market situation, and in many branches of insurance is known to exhibit strong
fluctuations; see Section 11.5 for a simple model. There have been attempts to
determine γ from theoretical arguments, see Young (2004) for a review, but these
efforts are not used much in practice and will not be considered.
The loading concept separates the market side from the insurance process itself,
but another issue is whether the pure premium is known. Stochastic models for X
always depend on unknown quantities such as parameters or probability distribu-
tions. They are determined from experience or even assessed informally if histor-
ical data are lacking, and there is a crucial distinction between the true πpu with
perfect knowledge of the underlying situation and the π̂pu used for analysis and
decisions. The discrepancy between what we seek and what we get is a fundamen-
tal issue of error that is present everywhere (see Figure 1.1), and there is special
notation for it. A parameter or quantity with a ˆ such as ψ̂ means an estimate or an
assessment of an underlying, unknown ψ. Chapter 7 offers a general discussion of
errors and how they are confronted.

1.2.3 Portfolios and solvency

A second major theme in insurance is control. Companies are obliged to set aside
funds to cover future obligations, and this is even a major theme in the legal defini-
tion of insurance. A company carries responsibility for many policies. It will lose
on some and gain on others. In property insurance policies without accidents are
profitable, those with large claims are not. Long lives in pension schemes lead to
losses, short ones to gains. At the portfolio level, gains and losses average out. This
is the beauty of a large agent handling many risks simultaneously.
Suppose a portfolio consists of J policies with claims X1 , . . . , X J . The total claim
is then
X = X1 + · · · + X J , (1.3)
where calligraphic letters like X will be used for quantities at the portfolio level.
We are certainly interested in E(X), but equally important is its distribution. Regu-
lators demand sufficient funds to cover X with high probability. The mathematical
formulation is in terms of a percentile q , which is the solution of the equation
Pr(X > q ) =  (1.4)

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page 6 — #32

i i

6 Introduction

where  is a small number (for example 1%). The amount q is known as the
solvency capital or reserve. Percentiles are used in finance too and are then often
called value at risk (or VaR for short). As elsewhere, the true q we seek is not the
same as the estimated q̂ we get.

1.2.4 Risk ceding and reinsurance

Risk is ceded from ordinary policy holders to companies, but companies do the
same thing between themselves. This is known as reinsurance, and the ceding
company is known as the cedent. The rationale could be the same; i.e., that a
financially weaker agent is passing risk to a stronger one. In reality even the largest
companies do this to diversify risk, and financially the cedent may be as strong
as the reinsurer. There is now a chain of responsibilities that can be depicted as
follows:
original clients −→ cedent −→ reinsurer
X (primary) Xce = X − Xre Xre (derived)
The original risk X is split between cedent and reinsurer through two separate
relationships, where the cedent part Xce is net and the difference between two cash
flows. Of course Xre ≤ X; i.e., the responsibility of the reinsurer is always less than
the original claim. Note the calligraphic style that applies to portfolios. There may
in practice be several rounds of such cedings in complicated networks extending
around the globe. One reinsurer may go to a second reinsurer, and so on. Modern
methods provide the means to analyse risk taken by an agent who is far away from
the primary source. Ceding and reinsurance are tools used by managers to tune
portfolios to a desired risk profile.

1.3 Financial risk: Basic concepts

1.3.1 Introduction
Gone are the days when insurance liabilities were insulated from assets and insur-
ance companies carried all the financial risk themselves. One trend is ceding to
customers. In countries like the USA and Britain, insurance products with financial
risk integrated have been sold for decades under names such as unit link or univer-
sal life. The rationale is that clients receive higher expected financial income in
exchange for carrying more risk. Pension plans today are increasingly contributed
benefits (CB), where financial risk rests with individuals. There is also much inter-
est in investment strategies tailored to given liabilities and how they distribute over
time. This is known as asset liability management (ALM) and is discussed in

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page 7 — #33

i i

1.3 Financial risk: Basic concepts 7

Chapter 15. The present section and the next one review the main concepts of
finance.

1.3.2 Rates of interest

An ordinary bank deposit v0 grows to (1 + r)v0 at the end of one period and to
(1 + r)K v0 after K of them. Here r, the rate of interest, depends on the length of
the period. Suppose interest is compounded over K segments, each of length 1/K,
so that the total time is one. If interest per segment is r/K, the value of the account
becomes
 r K
1+ v0 → er v0 , as K → ∞,
K
after one of the most famous limits of mathematics. Interest earnings may therefore
be cited as
rv0 or (er − 1)v0 ,
depending on whether we include ‘interest on interest’. The second form implies
continuous compounding of interest and higher earnings (er − 1 > r if r > 0), and
now (er )k = erk takes over from (1 + r)k . It doesn’t really matter which form we
choose, since they can be made equivalent by adjusting r.

1.3.3 Financial returns

Let V0 be the value of a financial asset at the start of a period and V1 the value at
the end of it. The relative gain
V 1 − V0
R= (1.5)
V0
is known as the return on the asset. Solving for V1 yields
V1 = (1 + R)V0 , (1.6)
with RV0 , the financial income. Clearly R acts like interest, but there is more to it
than that. Interest is a fixed benefit offered by a bank (or an issuer of a very secure
bond) in return for making a deposit and is risk free. Shares of company stock, in
contrast, are fraught with risk. They may go up (R positive) or down (R negative).
When dealing with such assets, V1 (and hence R) is determined by the market,
whereas with ordinary interest r is given and V1 follows.
The return R is the more general concept and is a random variable with a prob-
ability distribution. Take the randomness away, and we are back to a fixed rate of
interest r. As r depends on the time between V0 and V1 , so does the distribution of

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page 8 — #34

i i

8 Introduction

R; as will appear many times in this book. Whether the rate of interest r really is
risk free is not as obvious as it seems. True, you do get a fixed share of your deposit
as a reward, but that does not tell its worth in real terms when price increases are
taken into account. Indeed, over longer time horizons risk due to inflation may be
huge and even reduce the real value of cash deposits and bonds. Saving money with
a bank at a fixed rate of interest may also bring opportunity cost if the market rate
after a while exceeds what you get. These issues are discussed and integrated with
other sources of risk in Part III.

1.3.4 Log-returns
Economics and finance have often constructed stochastic models for R directly. An
alternative is the log-return

X = log(1 + R), (1.7)

which by (1.5) can be written X = log(V1 ) − log(V0 ); i.e., as a difference of log-

arithms. The modern theory of financial derivatives (Section 3.5 and Chapter 14)
is based on X. Actually, X and R do not necessarily deviate that strongly since the
Taylor series of log(1 + R) is

R 2 R3
X =R− + + ··· ,
2 3
where R (a fairly small number) dominates so that X  R, at least over short
periods. It follows that the distributions of R and X must be rather similar (see
Section 2.4), but this is not to say that the discrepancy is unimportant. It depends
on the amount of random variation present, and the longer the time horizon the
more X deviates from R; see Section 5.4 for an illustration.

1.3.5 Financial portfolios

Investments are often spread over many assets as baskets or financial portfolios.
By intuition this must reduce risk; see Section 5.3, where the issue is discussed.
A central quantity is the portfolio return, denoted R (in calligraphic style). Its rela-
tionship to the individual returns R j of the assets is as follows. Let V10 , . . . , V J0 be
investments in J assets. The portfolio value is then


J 
J
V0 = V j0 growing at the end of the period to V1 = (1 + R j )V j0 .
j=1 j=1

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page 9 — #35

i i

1.4 Risk over time 9

Subtract V0 from V1 and divide by V0 , and you get the portfolio return

J
V0 j
R= w jR j where w j = . (1.8)
j=1
V0

Here w j is the weight on asset j and

w1 + · · · + w J = 1. (1.9)

Financial weights define the distribution on individual assets and will, in this book,
usually be normalized so that they sum to 1.
The mathematics allow negative w j . With bank deposits this corresponds to bor-
rowing. It is also possible with shares, known as short selling. A loss due to a
negative development is then carried by somebody else. The mechanism is as fol-
lows. A short contract with a buyer is to sell shares at the end of the period at an
agreed price. At that point we shall have to buy at market price, gaining if it is lower
than our agreement, losing if not. Short contracts may be an instrument to lower
risk (see Section 5.3) and require liquidity; i.e., assets that are traded regularly.

1.4 Risk over time

1.4.1 Introduction
A huge number of problems in finance and insurance have time as one of the central
ingredients, and this requires additional quantities and concepts. Many of these are
introduced below. The emphasis is on finance, where the number of such quantities
is both more plentiful and more complex than in insurance. Time itself is worth a
comment. In this book it will be run on equidistant sequences, either

T k = kT or tk = kh
(1.10)
time scale for evaluation time scale for modelling

for k = 0, 1, . . . On the left, T is an accounting period (e.g., year, quarter, month) or

the time to expiry of a bond or an option. Financial returns Rk , portfolio values Vk
and insurance liabilities Xk are followed over T k . The present is always at T 0 = 0,
whereas k > 0 is the future which requires stochastic models to portray what is
likely and what is not. Negative time will sometimes be used for past values.
The time scale h is used for modelling. It may coincide with T , but it may well
be smaller so that T = Kh for K > 1. How models on different time scales are
related is important; see Section 5.7, where this issue is discussed. There is also
much scope for very short time increments where h → 0 (so that K = T/h → ∞).
This is known as continuous-time modelling and is above all a trick to find simple

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page 10 — #36

i i

10 Introduction

mathematical solutions. Parameters or variables are then often cited as intensities,

which are quantities per time unit. An example is interest rates, which will on
several occasions be designated rh with r an intensity and not a rate as in Section
1.3. Claim frequencies in property insurance (Chapter 8) and mortalities in life
insurance (Chapter 12) are other examples of using intensities for modelling. This
section is concerned with the macro time scale T only.

1.4.2 Accumulation of values

If v0 is the original value of a financial asset, by T K = KT it is worth

VK = (1 + R1 )(1 + R2 ) · · · (1 + RK )V0 = (1 + R0:K )v0 ,

where R1 , . . . , RK are the returns. This defines R0:K on the right as the K-step return
R0:K = (1 + R1 ) · · · (1 + RK ) − 1 and also X0:K = X1 + · · · + XK ,
(1.11)
ordinary returns log-returns

where Xk = log(1 + Rk ) and X0:K = log(1 + R0:K ). The log-returns on the right
are accumulated by adding them. Interest is a special case (an important one!) and
grows from T 0 = 0 to T K according to

r0:K = (1 + r1 )(1 + r2 ) · · · (1 + rK ) − 1, (1.12)

where r1 , . . . , rK are the future rates. This reduces to r0:K = (1 + r)K − 1 if all rk = r,
but in practice rk will float in a way that is unknown at T 0 = 0.
Often VK aggregates economic and financial activity beyond the initial invest-
ment v0 . Let Bk be income or expenses that surface at time T k , and suppose the
financial income or loss coming from the sequence B1 , . . . , BK is the same as for
the original asset. The total value at T K is then the sum

K
VK = (1 + R0:K )v0 + (1 + Rk:K )Bk , (1.13)
k=1

where Rk:K = (1 + Rk+1 ) · · · (1 + RK ) − 1 with RK:K = 0. Later in this section

B1 , . . . , BK will be a fixed cash flow, but further on (for example in Section 3.6)
there will be huge uncertainty as to what their values are going to be, with addi-
tional random variation on top of the financial variation.

1.4.3 Forward rates of interest

Future interest rates like rk or r0:K are hopeless to predict from mathematical mod-
els (you will see why in Section 6.4), but there is also a market view that conveys

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page 11 — #37

i i

1.4 Risk over time 11

the so-called implied rates. Consider an asset of value v0 that will be traded at time
T K for a price V0 (K) agreed today. Such contracts are called forwards and define
inherent rates of interest r0 (0 : K) through
V0 (K)
V0 (K) = {1 + r0 (0 : K)}v0 or r0 (0 : K) = − 1. (1.14)
v0
Note the difference from (1.12) where r0:K is uncertain, whereas now V0 (K) and
hence r0 (0 : K) is fixed by the contract and known at T 0 = 0. Forward rates can in
practice be deduced from many different sources, and the results are virtually iden-
tical. Otherwise there would have been market inconsistencies opening up money-
earning schemes with no risk attached; more on this in Chapter 14.
We are often interested in breaking the rate r0 (0 : K) down into its average value
r̄0 (0 : K) per period. The natural definition is
1 + r0 (0 : K) = {1 + r̄0 (0 : K)}K or r̄0 (0 : K) = {1 + r0 (0 : K)}1/K − 1, (1.15)
and as K is varied, the sequence r̄0 (0 : K) traces out the interest-rate curve or
yield curve, which is published daily in the financial press.

1.4.4 Present and fair values

What is the value today of receiving B1 at time T 1 ? Surely it must be B1 /(1 + r),
which grows to precisely B1 when interest is added. More generally, Bk at T k is
under constant rate of interest worth Bk /(1+r)k , today, which motivates the present
value (PV) as the value of a payment stream B0 , . . . , BK through

K
1
PV = dk Bk where dk = . (1.16)
k=0
(1 + r)k

This is a popular criterion in all spheres of economic life and used for assets and
liabilities alike. It applies even when B0 , . . . , BK are stochastic (then the present
value is too). Individual payments may be both positive and negative.
The quantities dk = 1/(1 + r)k (or dk = e−rk if interest is continuously com-
pounded) are known as discount factors; they devaluate or discount future income.
In life insurance, r is called the technical rate. The value to use isn’t obvious, least
of all with payment streams lasting decades ahead. Market discounting is an alter-
native. The coefficient dk in (1.16) is then replaced by
1 1
dk = = or dk = P0 (0 : k), (1.17)
{1 + r̄0 (0 : k)} k 1 + r0 (0 : k)
where P0 (0 : k) comes from the bond market; see below. Instead of choosing the
technical rate r administratively, the market view is used. The resulting valuation,

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page 12 — #38

i i

12 Introduction

known as fair value, holds obvious attraction, but also the disadvantage that such
discount sequences fluctuate randomly, bringing uncertainty even if there was none
at the start; see Section 15.3.

1.4.5 Bonds and yields

Governments and private companies raise capital by issuing bonds. In return for
money received upfront, the issuer makes fixed payments to bond holders at certain
points in time T 0 , . . . , T K . The last one at T K is a big one and known as the face
of the bond. Earlier payments can be seen as interest on a loan that size, but it is
simplest to define a bond as a fixed cash flow. How long it lasts varies enormously,
from a year or less up to half a century or even longer. Bonds have a huge second-
hand market and are traded regularly.
Should bonds be valued through present or fair values? Actually it is the other
way around. The present value is given by what the market is willing to pay, and the
rate of interest determined by the resulting equation. We are dealing with a fixed
payment stream B0 , . . . , BK . Let V0 be its market value at T 0 = 0. The yield y from
buying the rights to the stream is then the solution of the equation

K
Bk
V0 = . (1.18)
k=0
(1 + y)k

With more than one payment a numerical method such as bisection is needed to
determine y; see Appendix C.4.
A special case is the zero-coupon bond or T-bond, for which B0 = · · · = BK−1 =
0. It is unit faced if BK = 1. Now the only transaction occurs at maturity T K , and
in a market operating rationally its yield y is the same as the forward rate of interest
r̄0 (0 : K). The price of unit-faced T-bonds will be denoted P0 (0 : K), which is what
is charged today for the right to receive one money unit at T K and relates to the
forward rate of interest through
1 1
P0 (0 : K) = = . (1.19)
1 + r0 (0 : K) {1 + r̄0 (0 : K)}K
Why? Because anything else would have allowed riskless financial income, which
is incompatible with free markets. The prices P0 (0 : K) will be used a lot in Chap-
ters 14 and 15.

1.4.6 Duration
The timing of bonds and other fixed payment streams is often measured through
their duration D. There are several versions which vary in detail. One possibility is

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page 13 — #39

i i

1.4 Risk over time 13


K
Bk (1 + r)−k
D= qk T k where qk = K . (1.20)
−i
k=0 i=0 Bi (1 + r)

The sequence q0 , . . . , qK adds to one and is a probability distribution. Since qk is

proportional to the present value of the kth payment, D is formally an expectation
that expresses how long the cash flow B0 , . . . , BK is ‘on average’.
For a zero-coupon bond maturing at T K = KT , we have

qK = 1 and qk = 0, for k < K

so that D = T K , a sensible result! A bond with fixed coupon payments and a final
(much larger) face has duration between T K /2 and T K .

1.4.7 Investment strategies

Long-term management of financial risk is usually concerned with different classes
of assets which fluctuate jointly. Let Rk be the portfolio return in period k. The
account {Vk } then evolves according to

Vk = (1 + Rk )Vk−1 , k = 1, 2, . . . , (1.21)

where the link of Rk to the individual assets is through (1.8) as before. If R jk is the
return of asset j in period k, then


J
Rk = w j R jk , (1.22)
j=1

and the weights w1 , . . . , w J define the investment strategy. One way is to keep them
fixed, as in (1.22) where they do not depend on k. This is not achieved automati-
cally since individual investments meet with unequal success, which changes their
relative value. Weights can only be kept fixed by buying and selling. Restructuring
financial weights in such a way is known as rebalancing.
The opposite line is to let weights float freely. Mathematically this is more con-
veniently expressed through


J
Vk = Vk j where Vk j = (1 + Rk j )Vk−1, j , j = 1, . . . , J,
j=1

emphasizing assets rather than returns. For more on investment strategies, consult
Section 15.5.

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page 14 — #40

i i

14 Introduction

1.5 Method: A unified beginning

1.5.1 Introduction
How to make the preceding quantities and concepts flourish? Stochastic models
and Monte Carlo are needed! Here is an example introducting both. Let X1 , X2 , . . .
be an independent series driving Y1 , Y2 , . . . through the recursion

Yk = ak Yk−1 + Xk , k = 1, 2 . . . , starting at Y0 = y0 (1.23)

which covers a number of important situations. The series a1 , a2 , . . . may be fixed

coefficients. Another possibility is ak = 1 + r where r is a rate of interest, and now
Y1 , Y2 , . . . are values of an account influenced by random input. A more advanced
version is
ak = 1 + Rk and Xk = −Xk ,
(1.24)
financial risk insurance risk

and two different sources of risk that might themselves demand extensive mod-
elling and simulation are integrated. There will be more on this in Section 15.6.
Here the target is a more modest one. A simple Monte Carlo algorithm and nota-
tion for such schemes will first be presented and then four simple examples. The
aim is to introduce a general line of attack and indicate the power of Monte Carlo
for problem solving, learning and communication.

1.5.2 Monte Carlo algorithms and notation

Let Y1 , . . . , YK be the first K variables of the sequence (1.23). How they are simu-
lated is indicated by the schemes in Algorithm 1.1 (a skeleton!), which is the first
algorithm of the book. After initialization (Line 1) the random terms Xk∗ are drawn
(Line 3) and the new values Yk∗ found. All simulated variables are ∗ -marked, a con-
vention that will be followed everywhere. The backward arrow ← signifies that the
variable on the left is assigned the value on the right, more convenient notation
than an ordinary equality sign. For example, when only the last value YK∗ is wanted
(as is frequent), statements like Y ∗ ← aY ∗ + X ∗ simply overwrite Y ∗ , and values
of the past are not stored in the computer. The % symbol will be used for inserting
comments.
A huge number of simulation experiments in insurance and finance fit this
scheme or some simple variation of it, which suggests a fairly stable pattern of
Monte Carlo programming that can be lifted from one problem to another. Is K in
Algorithm 1.1 random (as in Example 2 below)? Draw it prior to entering the loop
on Line 2. Random ak as in (1.24)? Similarly, remove it from the input list and
generate it before computing Yk∗ on Line 4.

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page 15 — #41

i i

1.5 Method: A unified beginning 15

Algorithm 1.1 Basic recursion

0 Input: y0 , {ak }
1 Y0∗ ← y0 %Initialization
%Draw K here if random
2 For k = 1, . . . , K do
3 Sample Xk∗ %Many possibilities
4 Yk∗ ← ak Yk−1 ∗
+ Xk∗ %New value
5 Return Y0 , . . . , YK (or just YK∗ ).
∗ ∗

Annual claims Annual claims

0.04
100

Million US$
80

100 replications 0.03

60

0.02
40

0.01
20

0.0
0

0 2000 4000 6000 8000 10000 60 80 100 120

Number of policies simulated Million US$

Figure 1.2 Simulations of term insurance. Left: 100 parallel runs through insurance portfolio.
Right: Annual density function obtained from m = 10 000 simulations.

Example 1.5.3: Term insurance

Consider J contracts where the death of policy holders releases one-time payments.
How frequently this occurs is controlled by probabilities depending on age and sex
of the individuals, which would be available on file. The situation is so simple that
it is possible to examine portfolio liability and its uncertainty through mathematics
(this is done in Section 3.4), but the point now is to use Monte Carlo. Let X j be the
payoff for policy j, either 0 when the policy holder survives or s j when she (or he)
does not. The stochastic model is

Pr(X j = 0) = p j and Pr(X j = s j ) = 1 − p j ,

where p j is the probability of survival. A simulation goes through the entire port-
folio, reads policy information from file, draws those who die (whom we have to
pay for) and adds all the payments together. In Algorithm 1.1 take K = J, ak = 1
and X ∗j is either 0 or s j ; details in Section 3.4.
The example in Figure 1.2 shows annual expenses for J = 10 000 policies for
which s j = 1 for all j (money unit: one million US$). All policy holders were

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page 16 — #42

i i

16 Introduction

Density of total claim Density of total claim

1000 policies 100000 policies

0 50 100 150 2500 3000 3500 4000

Million DKK Million DKK

Figure 1.3 Density functions of the total claim against portfolio of fire risks (seven/eight Danish
kroner (DKK) for one euro).

between 30 and 60 years, with survival probabilities and age distribution as

specified in Section 3.4. One hundred parallel runs through the portfolio are
plotted jointly on the left, showing how the simulations evolve. The curved shape
has no significance. It is due to the age of the policy holders having been ordered
on file so that the young ones with low death rates are examined first. What counts
is the variation at the end, which is between 65 and 105 million. Another view is
provided by the probability density function on the right, which has been estimated
from the simulations by means of the kernel method in Section 2.2 (larger exper-
iment needed). The Gaussian shape follows from the Lindeberg extension of the
central limit theorem (Appendix A.4). Such risk is often ignored in life insurance,
but in this example uncertainty isn’t negligible.

Example 1.5.4: Property insurance

A classic model in property insurance is identical risks. Claims are then equally
likely for everybody, and losses exhibit no systematic variation between individu-
als. The portfolio payout is
X = Z1 + · · · + ZN ,
where N is the number of insurance incidents and Z1 , Z2 , . . . their cost. In Algo-
rithm 1.1 K = N (and drawn prior to the loop), ak = 1 and Xk = Zk ; consult
Algorithm 3.1 for details.
The example in Figure 1.3 was run with annual claim frequency 1% per policy
and a Poisson distribution for N. Losses Z1 , Z2 , . . . were drawn from the empirical
distribution function (Section 9.2) of the Danish fire data introduced in Section 9.6.
The latter is a record of more than 2000 industrial fires, the largest going up to
several hundred million Danish kroner (divide by seven or eight for euros, or ten
for dollars). Figure 1.3 shows the density function for the portfolio liabilities of a

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page 17 — #43

i i

1.5 Method: A unified beginning 17

Rapid change Slow change

15

15
% Interest % Interest
10

10
5

5
0

0 5 10 15 20 25 30 0 0 5 10 15 20 25 30
Years ahead Years ahead

Figure 1.4 Simulations of the annual rate of interest from the Vasic̆ek model.

‘small’ portfolio (J = 1000) on the left and a ‘large’ one (J = 100 000) on the right.
The uncertainty is now much larger than in the first example. Note that the density
function on the left is strongly skewed, but this asymmetry is straightened out as
the portfolio grows, and the distribution becomes more Gaussian. The central limit
theorem tells us this must be so.

Example 1.5.5: Reversion to mean

Monte Carlo is also a useful tool for examining the behaviour of stochastic models.
One type concerns interest rates, equity volatilities, rates of inflation and exchange
rates. All of these tend to fluctuate between certain, not clearly defined limits, and
if they move too far out on either side, there are forces in the economy that pull
them back again. This is known as reversion to mean and applies to a number of
financial variables.
Interest rates are an important example. One of the most popular models, pro-
posed by Vasic̆ek (1977), is the recursion

rk = Y k + ξ where Yk = aYk−1 + σεk , k = 1, 2, . . . , (1.25)

starting at Y0 = r0 −ξ. Here ξ, a and σ are fixed parameters and {εk } independent and
identically distributed variables with mean 0 and standard deviation 1. The model
is known as an autoregression of order 1 and is examined in Section 5.6. Here the
objective is simulation, which is carried out by taking ak = a in Algorithm 1.1 and
adding ξ to the output Y1∗ , Y2∗ . . . so that Monte Carlo interest rates r1∗ , r2∗ , . . . are
produced.

i i

i i
 i i

“9780521830485AR” — 2013/11/8 — 20:41 — page 18 — #44

i i

18 Introduction

Figure 1.4 shows simulated scenarios on an annual time scale under Gaussian
models. The parameters are

r0 = 3%, ξ = 7%, a = 0.70, σ = 0.016; r0 = 3%, ξ = 7%, a = 0.95, σ = 0.007

‘rapid’ change ‘slow’ change

both representing high-interest-rate regimes. The simulations start at r0 = 3%,

much lower than the long-term average ξ = 7%. That level is reached quickly on
the left with all traces of systematic changes in patterns having gone after about
5 years. Phenomena of this kind are known as stationary and are discussed in
Section 5.6. The same behaviour is observed with the second model on the right,
but now the transient period until fluctuations stabilize is much longer, and after 25
years movements may still be slightly on the rise upwards. This is caused by the
large value of a and is not realistic for interest rates in practice. Stationarity requires
−1 < a < 1 and model behaviour changes completely outside this interval, as will
emerge next.

Example 1.5.6: Equity over time

Stock prices {S k } are not mean reverting. They represent traded commodities, and
had it been possible to identify systematic factors that drove them up and down,
we would have been able to act upon them to earn money. But that opportunity
would have been available to everybody, removing the forces we were utilizing and
rendering the idea useless. Models for equity are for this reason very different from
those for interest rates, and based on stochastically independent returns. A common
specification is

Rk = eξ+σεk − 1, k = 1, 2, . . . , (1.26)

where εk is a sequence of independent random variables with mean 0 and standard

deviation 1. By definition S k = (1 + Rk )S k−1 so that

S k = S k−1 eξ+σεk , k = 1, 2, . . . , starting at S 0 = s0 . (1.27)

This is known as a geometric random walk with log(S k ) following an ordinary

random walk; see Section 5.5.
The behaviour of these models can be studied through Monte Carlo using Algo-
rithm 1.1. Apply it to Yk = log(S k ) (using ak = 1 and Xk = ξ + σεk ) and convert the
∗
simulations by taking S k∗ = eYk . The simulated scenarios in Figure 1.5 are monthly
and apply to the k-step returns R∗0:k = S k∗ /S 0 − 1 rather than the share price directly
(the initial value S 0 = s0 is then immaterial). Values for the parameters were:

i i

i i
Discovering Diverse Content Through
Random Scribd Documents
 I
Representation of Employees

1. Annual meetings for election of employee’s representatives.

Employees at each of the mining camps shall annually elect from

among their number representatives to act on their behalf with
respect to matters pertaining to their employment, working and
living conditions, the adjustment of differences, and such other
matters of mutual concern and interest as relations within the
industry may determine.

2. Time, place and method of calling annual meetings, and

persons entitled to be present and participate in the election of
representatives.

The annual meetings of employees for the election of their

representatives shall be held simultaneously at the several mining
camps on the second Saturday in January. The meetings shall be
called by direction of the president of the company. Notices of the
meetings, indicating their time and place, as well as the number of
representatives to be elected, shall be publicly posted at each camp
a week in advance, and shall state that employees being wage-
earners in the employ of the company at the time of the meeting
and for at least three months immediately preceding, but not
salaried employees, shall be entitled to be present and vote. Special
meetings shall be similarly called when removal, resignations, or
other circumstance occasions a vacancy in representation.

3. Method of conducting meetings, and reporting election of

representatives.
Each meeting for the election of employees’ representatives shall
choose its own chairman and secretary. At the appointed hour, the
meeting shall be called to order by one of the employees’
representatives, or, in the absence of a representative, any employee
present, and shall proceed to the election of a chairman and
secretary. The chairman shall conduct, and the secretary record, the
proceedings. They shall certify in writing to the president of the
company the names of the persons elected as the employees’
representatives for the ensuing year.

4. Basis and term of representation.

Representatives of employees in each camp shall be on the basis of

one representative to every one hundred and fifty wage-earners, but
each camp, whatever its number of employees, shall be entitled to
at least two representatives. Where the number of employees in any
one camp exceeds one hundred and fifty, or any multiple thereof, by
seventy-five or more, an additional representative shall be elected.
The persons elected shall act as the employees’ representatives from
the time of their election until the next annual meeting, unless in the
interval other representatives may, as above provided, have been
elected to take their places.

5. Nomination and election of representatives.

To facilitate the nomination and election of employees’

representatives, and to insure freedom of choice, both nomination
and election shall be by secret ballot, under conditions calculated to
insure an impartial count. The company shall provide ballot boxes
and blank ballots, differing in form, for purposes of nomination and
election. Upon entering the meeting, each employee entitled to be
present shall be given a nomination ballot on which he shall write
the names of the persons whom he desires to nominate as
representatives, and deposit the nomination ballot in the ballot box.
Each employee may nominate representatives to the number to
which the camp is entitled, and of which public notice has been
given. Employees unable to write may ask any of their fellow
employees to write for them on their ballots the names of the
persons whom they desire to nominate; but in the event of any
nomination paper containing more names than the number of
representatives to which the camp is entitled, the paper shall not be
counted. The persons—to the number of twice as many
representatives as the camp is entitled to—receiving the highest
number of nomination votes shall be regarded as the duly
nominated candidates for employees’ representatives, and shall be
voted upon as hereinafter provided. (For example: If a camp is
entitled to two representatives, the four persons receiving the
largest number of nominating votes shall be regarded as the duly
nominated candidates. If the camp is entitled to three
representatives, then the six persons receiving the largest number,
etc.)

6. Counting of nomination and election ballots.

The chairman shall appoint three tellers, who shall take charge of
the ballot box containing the nomination votes, and, with the aid of
the secretary, they shall make out the list of the duly nominated
candidates, which shall be announced by the chairman. The meeting
shall then proceed to elect representatives by secret ballot, from
among the number of candidates announced, the same tellers
having charge of the balloting. If dissatisfied with the count, either
as respects the nomination or election, any twenty-five employees
present may demand a recount, and for the purposes of the recount
the chairman shall select as tellers three from the number of those
demanding a recount, and himself assist in the counting, and these
four shall act, in making the recount, in place of the secretary and
the tellers previously chosen. There shall be no appeal from this
recount, except to the president of the company, and such appeal
may be taken as hereinafter provided, at the request of any twenty-
five employees present and entitled to vote.

7. Appeal in regard to nomination or election.

The chairman of the meeting shall preserve for a period of one week
both the nomination and election ballots. Should an appeal be made
to the president within seven days in regard to the validity of the
nomination or election, upon request in writing signed by twenty-five
employees present at the meeting, the chairman shall deliver the
ballots to the president of the company for recount. Should no such
request be received within that time, the chairman shall destroy the
ballots. If after considering the appeal the president is of the opinion
that the nomination or election has not been fairly conducted, he
shall order a new election at a time and place to be designated by
him.

8. General proceedings at meetings.

At annual meetings for the election of representatives, employees

may consider and make recommendations concerning any matters
pertaining to their employment, working or living conditions, or
arising out of existing industrial relations, including such as they may
desire to have their representatives discuss with the president and
officers of the company at the Annual Joint Conference of the
company’s officers and employees, also any matters referred to
them by the president, other officers of the company, the Advisory
Board or Social Joint Committee appointed at the preceding annual
joint conferences of officials and employees of the company. A
record of the proceedings shall be made by the secretary of the
meeting and certified to by the chairman, and copies delivered to
each of the representatives, to be retained by them for purposes of
future reference.
 II
District Conferences, Joint Committees and Joint
Meetings

1. District divisions.

To facilitate the purposes herein set forth, the camps of the company
shall be divided into five or more districts, as follows: the Trinidad
District, comprising all mines and coke oven plants in Las Animas
County; the Walsenburg District, comprising all mines in Huerfano
County; the Cañon District, comprising all mines in Fremont County;
the Western District, comprising all mines and coke oven plants
located on the Western Slope; the Sunrise District, comprising the
iron mines located in Wyoming.

2. Time, place and purpose of district conferences.

District conferences shall be held in each of the several districts

above mentioned at the call of the president, at places to be
designated by him, not later than two weeks following the annual
election of representatives, and at intervals of not more than four
months thereafter, as the operating officers of the company, or a
majority of the representatives of the employees in each of the
several districts, may find desirable. The purpose of these district
conferences shall be to discuss freely matters of mutual interest and
concern to the company and its employees, embracing a
consideration of suggestions to promote increased efficiency and
production, to improve working and living conditions, to enforce
discipline, avoid friction, and to further friendly and cordial relations
between the company’s officers and employees.

3. Representation at district conferences.

At the district conferences the company shall be represented by its
president or his representative and such other officials as the
president may designate. The employees shall be represented by
their elected representatives. The company’s representatives shall
not exceed in number the representatives of the employees. The
company shall provide at its own expense appropriate places of
meeting for the conferences.

4. Proceedings of district conferences.

The district conferences shall be presided over by the president of

the company, or such executive officer as he may designate. Each
conference shall select a secretary who shall record its proceedings.
The record of proceedings shall be certified to by the presiding
officer.

5. Joint committees on industrial relations.

The first district conferences held in each year shall select the
following joint committees on industrial relations for each district,
which joint committees shall be regarded as permanent committees
to be intrusted with such duties as are herein set forth, or as may be
assigned by the conferences. These joint committees shall be
available for consultation at any time throughout the year with the
Advisory Board on Social and Industrial Betterment, the president,
the president’s executive assistant, or any officer of the operating
department of the company.
(a) Joint Committee on Industrial Coöperation and Conciliation: to be
composed of six members.
(b) Joint Committee on Safety and Accidents: to be composed of six
members.
(c) Joint Committee on Sanitation, Health and Housing: to be
composed of six members.
(d) Joint Committee on Recreation and Education: to be composed
of six members.

6. Selection and composition of joint committees.

In selecting the members of the several joint committees on

industrial relations, the employees’ representatives shall, as respects
each committee, designate three members and the president of the
company or his representative, three members.

7. Duties of Joint Committee on Industrial Coöperation and

Conciliation.

The Joint Committee on Industrial Coöperation and Conciliation may,

of their own initiative, bring up for discussion at the joint
conferences, or have referred to them for consideration and report
to the president or other proper officer of the company at any time
throughout the year, any matter pertaining to the prevention and
settlement of industrial disputes, terms and conditions of
employment, maintenance of order and discipline in the several
camps, company stores, etc.

8. Duties of Joint Committee on Safety and Accidents.

The Joint Committee on Safety and Accidents may, of their own

initiative, bring up for discussion at the joint conferences, or have
referred to them for consideration and report to the president or
other proper officer of the company at any time throughout the year,
any matter pertaining to the inspection of mines, the prevention of
accidents, the safeguarding of machinery and dangerous working
places, the use of explosives, fire protection, first aid, etc.

9. Duties of Joint Committee on Sanitation, Health and Housing.

The Joint Committee on Sanitation, Health and Housing may, of their

own initiative, bring up for discussion at the joint conferences, or
have referred to them for consideration and report to the president
or other proper officer of the company at any time throughout the
year, any matter pertaining to health, hospitals, physicians, nurses,
occupational disease, tuberculosis, sanitation, water supply, sewage
system, garbage disposal, street cleaning, wash and locker rooms,
housing, homes, rents, gardens, fencing, etc.

10. Duties of Joint Committee on Recreation and Education.

The Joint Committee on Recreation and Education may, of their own

initiative, bring up for discussion at the joint conferences, or have
referred to them for consideration and report to the president or
other proper officer of the company, at any time throughout the
year, any matter pertaining to social centers, club houses, halls,
playgrounds, entertainments, moving pictures, athletics,
competitions, field days, holidays, schools, libraries, classes for those
who speak only foreign languages, technical education, manual
training, health lectures, classes in first aid, religious exercises,
churches and Sunday schools, Y. M. C. A. organizations, etc.

11. Annual and special joint meetings.

In addition to the district conferences in each of the several districts,

there shall be held in the month of December an annual joint
meeting, at a time and place to be designated by the president of
the company, to be attended by the president and such officers of
the company as he may select and by all the employees’
representatives of the several districts. At this meeting reports
covering the work of the year shall be made by the several joint
committees and matters of common interest requiring collective
action considered. A special joint meeting of any two or more
districts may be called at any time upon the written request to the
president of a majority of the representatives in such districts or
upon the president’s own initiative, for the consideration of such
matters of common interest as cannot be dealt with satisfactorily at
district conferences. Notice of such special joint meetings shall be
given at least two weeks in advance.
 III
The Prevention and Adjustment of Industrial
Disputes

1. Observance of laws, rides and regulations.

There shall be on the part of the company and its employees, a strict
observance of the Federal and State laws respecting mining and
labor and of the company’s rules and regulations supplementing the
same.

2. Posting of wages and rules.

The scale of wages and the rules in regard to working conditions

shall be posted in a conspicuous place at or near every mine.

3. No discrimination on account of membership or non-

membership in labor or other organizations.

There shall be no discrimination by the company or by any of its

employees on account of membership or non-membership in any
society, fraternity or union.

4. The right to hire and discharge, and the management of the

properties.

The right to hire and discharge, the management of the properties,

and the direction of the working forces, shall be vested exclusively in
the company, and, except as expressly restricted, this right shall not
be abridged by anything contained herein.

5. Employees’ right to caution or suspension before discharge.

There shall be posted at each property a list of offenses for
commission of which by an employee dismissal may result without
notice. For other offenses, employees shall not be discharged
without first having been notified that a repetition of the offense will
be cause for dismissal. A copy of this notification shall, at the time of
its being given to an employee, be sent also to the president’s
industrial representative and retained by him for purposes of future
reference. Nothing herein shall abridge the right of the company to
relieve employees from duty because of lack of work. Where relief
from duty through lack of work becomes necessary, men with
families shall, all things being equal, be given preference.

6. Employees’ right to hold meetings.

Employees shall have the right to hold meetings at appropriate

places on company property or elsewhere as they may desire
outside of working hours or on idle days.

7. Employees’ right to purchase where they please.

Employees shall not be obliged to trade at the company stores, but

shall be at perfect liberty to purchase goods wherever they may
choose to do so.

8. Employees’ right to employ checkweighmen.

As provided by statute, miners have the right to employ

checkweighmen, and the company shall grant the said
checkweighmen every facility to enable them to render a correct
account of all coal weighed.

9. Employees’ right of appeal to president of company against

unfair conditions or treatment.

Subject to the provisions hereinafter mentioned, every employee

shall have the right of ultimate appeal to the president of the
company concerning any condition or treatment to which he may be
subjected and which he may deem unfair.

10. Duty of president’s industrial representative.

It shall be the duty of the president’s industrial representative to

respond promptly to any request from employees’ representatives
for his presence at any of the camps and to visit all of them as often
as possible, but not less frequently than once every three months, to
confer with the employees or their representatives and the
superintendents respecting working and living conditions, the
observance of Federal and State laws, the carrying out of company
regulations, and to report the result of such conferences to the
president.

11. Complaints and grievances to be taken up first with foremen

and superintendents.

Before presenting any grievance to the president, the president’s

industrial representative, or other of the higher officers of the
company, employees shall first seek to have differences or the
conditions complained about adjusted by conference, in person or
through their representatives, with the mine superintendent.

12. Investigation of grievances by president’s industrial

representative.

Employees believing themselves to be subjected to unfair conditions

or treatment and having failed to secure satisfactory adjustment of
the same through the mine superintendent may present their
grievances to the president’s industrial representative, either in
person or through their regularly elected representatives, and it shall
be the duty of the president’s industrial representative to look into
the same immediately and seek to adjust the grievance.

13. The right of appeal to the superior officers of the company

against unfair treatment, conditions, suspensions or dismissals.
Should the president’s industrial representative fail to satisfactorily
conciliate any difference, with respect to any grievance, suspension
or dismissal, the aggrieved employee, either himself or through his
representative—and in either case in person or by letter—may
appeal for the consideration and adjustment of his grievance to the
division superintendent, assistant manager or manager, general
manager or the president of the company, in consecutive order. To
entitle an employee to the consideration of his appeal by any of the
higher officers herein mentioned, the right to appeal must be
exercised within a period of two weeks after the same has been
referred to the president’s industrial representative without
satisfactory redress.

14. Reference of differences in certain cases to Joint

Committees on Industrial Coöperation and Conciliation.

Where the president’s industrial representative or one of the higher

officials of the company fails to adjust a difference satisfactorily,
upon request to the president by the employees’ representatives or
upon the initiative of the president himself, the difference shall be
referred to the Joint Committee on Industrial Coöperation and
Conciliation of the district and the decision of the majority of such
joint committee shall be binding upon all parties.

15. Representation on joint committees to be equal when

considering adjustment of differences.

Whenever a Joint Committee on Industrial Coöperation and

Conciliation is called upon to act with reference to any difference,
except by the consent of all present the joint committee shall not
proceed with any important part of its duties unless both sides are
equally represented. Where agreeable, equal representation may be
effected by the withdrawal of one or more members from the side of
the joint committee having the majority.

16. Umpire to act with joint committees in certain cases.

Should the Joint Committee on Industrial Coöperation and
Conciliation to which a difference may have been referred fail to
reach a majority decision in respect thereto, if a majority of its
members so agree, the joint committee may select as umpire a third
person who shall sit in conference with the committee and whose
decision shall be binding upon all parties.

17. Arbitration or investigation in certain cases.

In the event of the Joint Committee on Industrial Coöperation and

Conciliation failing satisfactorily to adjust a difference by a majority
decision or by agreement on the selection of an umpire, as
aforementioned, within ten days of a report to the president of the
failure of the joint committee to adjust the difference, if the parties
so agree, the matter shall be referred to arbitration, otherwise it
shall be made the subject of investigation by the State of Colorado
Industrial Commission, in accordance with the provisions of the
statute regulating the powers of the commission in this particular.
Where a difference is referred to arbitration, one person shall be
selected as arbitrator if the parties can agree upon his selection.
Otherwise there shall be a board of three arbitrators, one to be
selected by the employees’ representatives on the Joint Committee
of Industrial Coöperation and Conciliation in the district in which the
dispute arises, one by the company’s representatives on this
committee, and a third by the two arbitrators thus selected.
By consent of the members of the Joint Committee on Industrial
Coöperation and Conciliation to which a difference has been
referred, the Industrial Commission of the State of Colorado may be
asked to appoint all of the arbitrators or itself arbitrate the
difference. The decision of the sole arbitrator or of the majority of
the Board of Arbitration or of the members of the State of Colorado
Industrial Commission when acting as arbitrators, as the case may
be, shall be final and shall be binding upon the parties.
 18. Protection of employees’ representatives against
discrimination.

To protect against the possibility of unjust treatment because of any

action taken or to be taken by them on behalf of one or more of the
company’s employees, any employees’ representative believing
himself to be discriminated against for such a cause shall have the
same right of appeal to the officers of the company or to the Joint
Committee on Industrial Coöperation and Conciliation in his district
as is accorded every other employee of the company. Having
exercised this right in the consecutive order indicated without
obtaining satisfaction, for thirty days thereafter he shall have the
further right of appeal to the Industrial Commission of the State of
Colorado, which body shall determine whether or not discrimination
has been shown, and as respects any representative deemed by the
Commission to have been unfairly dealt with, the company shall
make such reparation as the State of Colorado Industrial
Commission may deem just.
 IV
Social and Industrial Betterment

1. Executive supervision.

The president’s executive assistant, in addition to other duties, shall,

on behalf of the president, supervise the administration of the
company’s policies respecting social and industrial betterment.

2. Coöperation of president’s executive assistant with joint

committees in carrying out policies of social and industrial
betterment.

In the discharge of his duties, the president’s executive assistant

shall from time to time confer with the several Joint Committees, on
Industrial Coöperation and Conciliation, on Safety and Accidents, on
Sanitation, Health and Housing, and on Recreation and Education,
appointed at the annual joint conferences, as to improvements or
changes likely to be of mutual advantage to the company and its
employees. Members of the several joint committees shall be at
liberty to communicate at any time with the president’s executive
assistant with respect to any matters under their observation or
brought to their attention by employees or officials of the company,
which they believe should be looked into or changed. As far as may
be possible, employees should be made to feel that the president’s
executive assistant will welcome conferences with members of the
several joint committees on matters of concern to the employees,
whenever such matters have a direct bearing on the industrial,
social, and moral well-being of employees and their families or the
communities in which they reside.

3. Advisory Board on Social and Industrial Betterment.

In addition to consulting, from time to time, the several joint
committees or their individual members, the president’s executive
assistant shall be the chairman of a permanent Advisory Board on
Social and Industrial Betterment, to which may be referred questions
of policy respecting social and industrial betterment and related
matters requiring executive action.

4. Members of Advisory Board.

The Advisory Board on Social and Industrial Betterment shall be

composed of such of the company’s officers as the president may
designate.

5. Regular and special meetings of Advisory Board.

The Advisory Board shall meet at least once in every six months, and
may convene for special meetings upon the call of the chairman
whenever he may deem a special meeting advisable.

6. Powers and duties of the Advisory Board.

The Advisory Board shall have power to consider all matters referred
to it by the chairman, or any of its members, or by any committee or
organization directly or indirectly connected with the company, and
may make such recommendations to the president as in its opinion
seem to be expedient and in the interest of the company and its
employees.

7. Supervision of community needs by president’s executive

assistant.

The president’s executive assistant shall also exercise a general

supervision over the sanitary, medical, educational, religious, social
and other like needs of the different industrial communities, with a
view of seeing that such needs are suitably and adequately provided
for, and the several activities pertaining thereto harmoniously
conducted.
 8. Method of carrying out improvements.

Improvements respecting social and industrial betterment shall, after

approval by the president, be carried out through the regular
company organization.

9. Hospitals and doctors.

In camps where arrangements for doctors and hospitals have

already been made and are satisfactory, such arrangements shall
continue.
In making any new arrangement for a doctor, the employees’
representatives in the camps concerned, the president’s executive
assistant, and the chief medical officer shall select a doctor, and
enter into an agreement with him which shall be signed by all four
parties.

10. Company periodical.

The company shall publish, under the direction of the president’s

executive assistant, a periodical which shall be a means of
communication between the management, the employees and the
public, concerning the policies and activities of the company. This
periodical shall be used as a means of coördinating, harmonizing,
and furthering the social and industrial betterment work, and of
informing employees of the personnel and proceedings of
conferences, boards and committees in which they are interested. It
shall record events pertaining to social and industrial activities, and
be a medium for making announcements with reference to the
same, and for diffusing information of mutual interest to the
company and its employees.

11. Cost of administering plan of representation and of

furthering social and industrial betterment policies.
The promotion of harmony and good-will between the company and
its employees and the furtherance of the well-being of employees
and their families and the communities in which they reside being
essential to the successful operation of the company’s industries in
an enlightened and profitable manner, the expenses necessarily
incident to the carrying out of the social and industrial betterment
policies herein described, and the plan of representation, joint
conferences and joint meetings, herein set forth, including the
payment of traveling expenses of employees’ representatives when
attending joint conferences and annual joint meetings, and their
reimbursement for the working time necessarily lost in so doing,
shall be borne by the company. But nothing herein shall preclude
employees of the company from making such payment to their
representatives in consideration of services rendered on their behalf
as they themselves may voluntarily desire and agree to make.
 *** END OF THE PROJECT GUTENBERG EBOOK THE PERSONAL
RELATION IN INDUSTRY ***

Updated editions will replace the previous one—the old editions will
be renamed.

Creating the works from print editions not protected by U.S.

copyright law means that no one owns a United States copyright in
these works, so the Foundation (and you!) can copy and distribute it
in the United States without permission and without paying
copyright royalties. Special rules, set forth in the General Terms of
Use part of this license, apply to copying and distributing Project
Gutenberg™ electronic works to protect the PROJECT GUTENBERG™
concept and trademark. Project Gutenberg is a registered trademark,
and may not be used if you charge for an eBook, except by following
the terms of the trademark license, including paying royalties for use
of the Project Gutenberg trademark. If you do not charge anything
for copies of this eBook, complying with the trademark license is
very easy. You may use this eBook for nearly any purpose such as
creation of derivative works, reports, performances and research.
Project Gutenberg eBooks may be modified and printed and given
away—you may do practically ANYTHING in the United States with
eBooks not protected by U.S. copyright law. Redistribution is subject
to the trademark license, especially commercial redistribution.

START: FULL LICENSE

THE FULL PROJECT GUTENBERG LICENSE
 PLEASE READ THIS BEFORE YOU DISTRIBUTE OR USE THIS WORK

To protect the Project Gutenberg™ mission of promoting the free

distribution of electronic works, by using or distributing this work (or
any other work associated in any way with the phrase “Project
Gutenberg”), you agree to comply with all the terms of the Full
Project Gutenberg™ License available with this file or online at
www.gutenberg.org/license.

Section 1. General Terms of Use and

Redistributing Project Gutenberg™
electronic works
1.A. By reading or using any part of this Project Gutenberg™
electronic work, you indicate that you have read, understand, agree
to and accept all the terms of this license and intellectual property
(trademark/copyright) agreement. If you do not agree to abide by all
the terms of this agreement, you must cease using and return or
destroy all copies of Project Gutenberg™ electronic works in your
possession. If you paid a fee for obtaining a copy of or access to a
Project Gutenberg™ electronic work and you do not agree to be
bound by the terms of this agreement, you may obtain a refund
from the person or entity to whom you paid the fee as set forth in
paragraph 1.E.8.

1.B. “Project Gutenberg” is a registered trademark. It may only be

used on or associated in any way with an electronic work by people
who agree to be bound by the terms of this agreement. There are a
few things that you can do with most Project Gutenberg™ electronic
works even without complying with the full terms of this agreement.
See paragraph 1.C below. There are a lot of things you can do with
Project Gutenberg™ electronic works if you follow the terms of this
agreement and help preserve free future access to Project
Gutenberg™ electronic works. See paragraph 1.E below.
1.C. The Project Gutenberg Literary Archive Foundation (“the
Foundation” or PGLAF), owns a compilation copyright in the
collection of Project Gutenberg™ electronic works. Nearly all the
individual works in the collection are in the public domain in the
United States. If an individual work is unprotected by copyright law
in the United States and you are located in the United States, we do
not claim a right to prevent you from copying, distributing,
performing, displaying or creating derivative works based on the
work as long as all references to Project Gutenberg are removed. Of
course, we hope that you will support the Project Gutenberg™
mission of promoting free access to electronic works by freely
sharing Project Gutenberg™ works in compliance with the terms of
this agreement for keeping the Project Gutenberg™ name associated
with the work. You can easily comply with the terms of this
agreement by keeping this work in the same format with its attached
full Project Gutenberg™ License when you share it without charge
with others.

1.D. The copyright laws of the place where you are located also
govern what you can do with this work. Copyright laws in most
countries are in a constant state of change. If you are outside the
United States, check the laws of your country in addition to the
terms of this agreement before downloading, copying, displaying,
performing, distributing or creating derivative works based on this
work or any other Project Gutenberg™ work. The Foundation makes
no representations concerning the copyright status of any work in
any country other than the United States.

1.E. Unless you have removed all references to Project Gutenberg:

1.E.1. The following sentence, with active links to, or other

immediate access to, the full Project Gutenberg™ License must
appear prominently whenever any copy of a Project Gutenberg™
work (any work on which the phrase “Project Gutenberg” appears,
or with which the phrase “Project Gutenberg” is associated) is
accessed, displayed, performed, viewed, copied or distributed:
 This eBook is for the use of anyone anywhere in the United
States and most other parts of the world at no cost and with
almost no restrictions whatsoever. You may copy it, give it away
or re-use it under the terms of the Project Gutenberg License
included with this eBook or online at www.gutenberg.org. If you
are not located in the United States, you will have to check the
laws of the country where you are located before using this
eBook.

1.E.2. If an individual Project Gutenberg™ electronic work is derived

from texts not protected by U.S. copyright law (does not contain a
notice indicating that it is posted with permission of the copyright
holder), the work can be copied and distributed to anyone in the
United States without paying any fees or charges. If you are
redistributing or providing access to a work with the phrase “Project
Gutenberg” associated with or appearing on the work, you must
comply either with the requirements of paragraphs 1.E.1 through
1.E.7 or obtain permission for the use of the work and the Project
Gutenberg™ trademark as set forth in paragraphs 1.E.8 or 1.E.9.

1.E.3. If an individual Project Gutenberg™ electronic work is posted

with the permission of the copyright holder, your use and distribution
must comply with both paragraphs 1.E.1 through 1.E.7 and any
additional terms imposed by the copyright holder. Additional terms
will be linked to the Project Gutenberg™ License for all works posted
with the permission of the copyright holder found at the beginning
of this work.

1.E.4. Do not unlink or detach or remove the full Project

Gutenberg™ License terms from this work, or any files containing a
part of this work or any other work associated with Project
Gutenberg™.

1.E.5. Do not copy, display, perform, distribute or redistribute this

electronic work, or any part of this electronic work, without
prominently displaying the sentence set forth in paragraph 1.E.1
with active links or immediate access to the full terms of the Project
Gutenberg™ License.

1.E.6. You may convert to and distribute this work in any binary,
compressed, marked up, nonproprietary or proprietary form,
including any word processing or hypertext form. However, if you
provide access to or distribute copies of a Project Gutenberg™ work
in a format other than “Plain Vanilla ASCII” or other format used in
the official version posted on the official Project Gutenberg™ website
(www.gutenberg.org), you must, at no additional cost, fee or
expense to the user, provide a copy, a means of exporting a copy, or
a means of obtaining a copy upon request, of the work in its original
“Plain Vanilla ASCII” or other form. Any alternate format must
include the full Project Gutenberg™ License as specified in
paragraph 1.E.1.

1.E.7. Do not charge a fee for access to, viewing, displaying,

performing, copying or distributing any Project Gutenberg™ works
unless you comply with paragraph 1.E.8 or 1.E.9.

1.E.8. You may charge a reasonable fee for copies of or providing

access to or distributing Project Gutenberg™ electronic works
provided that:

• You pay a royalty fee of 20% of the gross profits you derive
from the use of Project Gutenberg™ works calculated using the
method you already use to calculate your applicable taxes. The
fee is owed to the owner of the Project Gutenberg™ trademark,
but he has agreed to donate royalties under this paragraph to
the Project Gutenberg Literary Archive Foundation. Royalty
payments must be paid within 60 days following each date on
which you prepare (or are legally required to prepare) your
periodic tax returns. Royalty payments should be clearly marked
as such and sent to the Project Gutenberg Literary Archive
Foundation at the address specified in Section 4, “Information
 about donations to the Project Gutenberg Literary Archive
Foundation.”

• You provide a full refund of any money paid by a user who

notifies you in writing (or by e-mail) within 30 days of receipt
that s/he does not agree to the terms of the full Project
Gutenberg™ License. You must require such a user to return or
destroy all copies of the works possessed in a physical medium
and discontinue all use of and all access to other copies of
Project Gutenberg™ works.

• You provide, in accordance with paragraph 1.F.3, a full refund of

any money paid for a work or a replacement copy, if a defect in
the electronic work is discovered and reported to you within 90
days of receipt of the work.

• You comply with all other terms of this agreement for free
distribution of Project Gutenberg™ works.

1.E.9. If you wish to charge a fee or distribute a Project Gutenberg™

electronic work or group of works on different terms than are set
forth in this agreement, you must obtain permission in writing from
the Project Gutenberg Literary Archive Foundation, the manager of
the Project Gutenberg™ trademark. Contact the Foundation as set
forth in Section 3 below.

1.F.

1.F.1. Project Gutenberg volunteers and employees expend

considerable effort to identify, do copyright research on, transcribe
and proofread works not protected by U.S. copyright law in creating
the Project Gutenberg™ collection. Despite these efforts, Project
Gutenberg™ electronic works, and the medium on which they may
be stored, may contain “Defects,” such as, but not limited to,
incomplete, inaccurate or corrupt data, transcription errors, a
copyright or other intellectual property infringement, a defective or
damaged disk or other medium, a computer virus, or computer
codes that damage or cannot be read by your equipment.

1.F.2. LIMITED WARRANTY, DISCLAIMER OF DAMAGES - Except for

the “Right of Replacement or Refund” described in paragraph 1.F.3,
the Project Gutenberg Literary Archive Foundation, the owner of the
Project Gutenberg™ trademark, and any other party distributing a
Project Gutenberg™ electronic work under this agreement, disclaim
all liability to you for damages, costs and expenses, including legal
fees. YOU AGREE THAT YOU HAVE NO REMEDIES FOR
NEGLIGENCE, STRICT LIABILITY, BREACH OF WARRANTY OR
BREACH OF CONTRACT EXCEPT THOSE PROVIDED IN PARAGRAPH
1.F.3. YOU AGREE THAT THE FOUNDATION, THE TRADEMARK
OWNER, AND ANY DISTRIBUTOR UNDER THIS AGREEMENT WILL
NOT BE LIABLE TO YOU FOR ACTUAL, DIRECT, INDIRECT,
CONSEQUENTIAL, PUNITIVE OR INCIDENTAL DAMAGES EVEN IF
YOU GIVE NOTICE OF THE POSSIBILITY OF SUCH DAMAGE.

1.F.3. LIMITED RIGHT OF REPLACEMENT OR REFUND - If you

discover a defect in this electronic work within 90 days of receiving
it, you can receive a refund of the money (if any) you paid for it by
sending a written explanation to the person you received the work
from. If you received the work on a physical medium, you must
return the medium with your written explanation. The person or
entity that provided you with the defective work may elect to provide
a replacement copy in lieu of a refund. If you received the work
electronically, the person or entity providing it to you may choose to
give you a second opportunity to receive the work electronically in
lieu of a refund. If the second copy is also defective, you may
demand a refund in writing without further opportunities to fix the
problem.

1.F.4. Except for the limited right of replacement or refund set forth
in paragraph 1.F.3, this work is provided to you ‘AS-IS’, WITH NO
OTHER WARRANTIES OF ANY KIND, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO WARRANTIES OF
MERCHANTABILITY OR FITNESS FOR ANY PURPOSE.

1.F.5. Some states do not allow disclaimers of certain implied

warranties or the exclusion or limitation of certain types of damages.
If any disclaimer or limitation set forth in this agreement violates the
law of the state applicable to this agreement, the agreement shall be
interpreted to make the maximum disclaimer or limitation permitted
by the applicable state law. The invalidity or unenforceability of any
provision of this agreement shall not void the remaining provisions.

1.F.6. INDEMNITY - You agree to indemnify and hold the Foundation,

the trademark owner, any agent or employee of the Foundation,
anyone providing copies of Project Gutenberg™ electronic works in
accordance with this agreement, and any volunteers associated with
the production, promotion and distribution of Project Gutenberg™
electronic works, harmless from all liability, costs and expenses,
including legal fees, that arise directly or indirectly from any of the
following which you do or cause to occur: (a) distribution of this or
any Project Gutenberg™ work, (b) alteration, modification, or
additions or deletions to any Project Gutenberg™ work, and (c) any
Defect you cause.

Section 2. Information about the Mission

of Project Gutenberg™
Project Gutenberg™ is synonymous with the free distribution of
electronic works in formats readable by the widest variety of
computers including obsolete, old, middle-aged and new computers.
It exists because of the efforts of hundreds of volunteers and
donations from people in all walks of life.

Volunteers and financial support to provide volunteers with the

assistance they need are critical to reaching Project Gutenberg™’s
goals and ensuring that the Project Gutenberg™ collection will
remain freely available for generations to come. In 2001, the Project
Gutenberg Literary Archive Foundation was created to provide a
secure and permanent future for Project Gutenberg™ and future
generations. To learn more about the Project Gutenberg Literary
Archive Foundation and how your efforts and donations can help,
see Sections 3 and 4 and the Foundation information page at
www.gutenberg.org.

Section 3. Information about the Project

Gutenberg Literary Archive Foundation
The Project Gutenberg Literary Archive Foundation is a non-profit
501(c)(3) educational corporation organized under the laws of the
state of Mississippi and granted tax exempt status by the Internal
Revenue Service. The Foundation’s EIN or federal tax identification
number is 64-6221541. Contributions to the Project Gutenberg
Literary Archive Foundation are tax deductible to the full extent
permitted by U.S. federal laws and your state’s laws.

The Foundation’s business office is located at 809 North 1500 West,

Salt Lake City, UT 84116, (801) 596-1887. Email contact links and up
to date contact information can be found at the Foundation’s website
and official page at www.gutenberg.org/contact

Section 4. Information about Donations to

the Project Gutenberg Literary Archive
Foundation
Project Gutenberg™ depends upon and cannot survive without
widespread public support and donations to carry out its mission of
increasing the number of public domain and licensed works that can
be freely distributed in machine-readable form accessible by the
widest array of equipment including outdated equipment. Many
small donations ($1 to $5,000) are particularly important to
maintaining tax exempt status with the IRS.

The Foundation is committed to complying with the laws regulating

charities and charitable donations in all 50 states of the United
States. Compliance requirements are not uniform and it takes a
considerable effort, much paperwork and many fees to meet and
keep up with these requirements. We do not solicit donations in
locations where we have not received written confirmation of
compliance. To SEND DONATIONS or determine the status of
compliance for any particular state visit www.gutenberg.org/donate.

While we cannot and do not solicit contributions from states where

we have not met the solicitation requirements, we know of no
prohibition against accepting unsolicited donations from donors in
such states who approach us with offers to donate.

International donations are gratefully accepted, but we cannot make

any statements concerning tax treatment of donations received from
outside the United States. U.S. laws alone swamp our small staff.

Please check the Project Gutenberg web pages for current donation
methods and addresses. Donations are accepted in a number of
other ways including checks, online payments and credit card
donations. To donate, please visit: www.gutenberg.org/donate.

Section 5. General Information About

Project Gutenberg™ electronic works
Professor Michael S. Hart was the originator of the Project
Gutenberg™ concept of a library of electronic works that could be
freely shared with anyone. For forty years, he produced and
distributed Project Gutenberg™ eBooks with only a loose network of
volunteer support.
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

ebookultra.com